init

.gitattributes

@@ -1,3 +1,5 @@
+*.jsonl filter=lfs diff=lfs merge=lfs -text
+*.json filter=lfs diff=lfs merge=lfs -text
 *.7z filter=lfs diff=lfs merge=lfs -text
 *.arrow filter=lfs diff=lfs merge=lfs -text
 *.bin filter=lfs diff=lfs merge=lfs -text
@@ -53,3 +55,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.jpg filter=lfs diff=lfs merge=lfs -text
 *.jpeg filter=lfs diff=lfs merge=lfs -text
 *.webp filter=lfs diff=lfs merge=lfs -text
+*.spacy filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1 @@
+training/

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2024 William Mattingly
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -1,3 +1,51 @@
 ---
 license: mit
 ---
+
+# Overall Model Performance
+| Model       |   Precision |   Recall |   F-Score |
+|:------------|------------:|---------:|----------:|
+| Small       |        94.1 |     89.2 |      91.6 |
+| Medium      |        94   |     90.5 |      92.2 |
+| Large       |        94.1 |     91.7 |      92.9 |
+| Transformer |        93.6 |     91.6 |      92.6 |
+
+# Performance per Label
+| Model       | Label           |   Precision |   Recall |   F-Score |
+|:------------|:----------------|------------:|---------:|----------:|
+| Small       | BUILDING        |        94.7 |     90.2 |      92.4 |
+| Medium      | BUILDING        |        95.2 |     92.8 |      94   |
+| Large       | BUILDING        |        94.8 |     93.2 |      94   |
+| Transformer | BUILDING        |        94.3 |     94.2 |      94.3 |
+| Small       | COUNTRY         |        97.6 |     94.6 |      96.1 |
+| Medium      | COUNTRY         |        96.5 |     96.3 |      96.4 |
+| Large       | COUNTRY         |        97.7 |     96.8 |      97.2 |
+| Transformer | COUNTRY         |        96.6 |     96.8 |      96.7 |
+| Small       | DLF             |        92.4 |     86.4 |      89.3 |
+| Medium      | DLF             |        95   |     84.1 |      89.2 |
+| Large       | DLF             |        93.5 |     88.4 |      90.9 |
+| Transformer | DLF             |        94.1 |     90.4 |      92.2 |
+| Small       | ENV_FEATURES    |        86.6 |     81.2 |      83.8 |
+| Medium      | ENV_FEATURES    |        86.3 |     79.1 |      82.5 |
+| Large       | ENV_FEATURES    |        77.5 |     90.1 |      83.3 |
+| Transformer | ENV_FEATURES    |        85.1 |     86.9 |      86   |
+| Small       | INT_SPACE       |        93.8 |     85.9 |      89.6 |
+| Medium      | INT_SPACE       |        93.9 |     91.3 |      92.6 |
+| Large       | INT_SPACE       |        92.4 |     93.8 |      93.1 |
+| Transformer | INT_SPACE       |        94.6 |     91.8 |      93.2 |
+| Small       | NPIP            |        92.7 |     86.4 |      89.4 |
+| Medium      | NPIP            |        94.5 |     82.4 |      88   |
+| Large       | NPIP            |        92.7 |     86.6 |      89.6 |
+| Transformer | NPIP            |        94.8 |     83   |      88.5 |
+| Small       | POPULATED_PLACE |        94   |     90.6 |      92.3 |
+| Medium      | POPULATED_PLACE |        93   |     91.2 |      92.1 |
+| Large       | POPULATED_PLACE |        95.2 |     90.4 |      92.7 |
+| Transformer | POPULATED_PLACE |        92.1 |     91.3 |      91.7 |
+| Small       | REGION          |        84.4 |     68.4 |      75.6 |
+| Medium      | REGION          |        81.4 |     75.8 |      78.5 |
+| Large       | REGION          |        83   |     76.8 |      79.8 |
+| Transformer | REGION          |        81.2 |     68.4 |      74.3 |
+| Small       | SPATIAL_OBJ     |        96   |     90   |      92.9 |
+| Medium      | SPATIAL_OBJ     |        95.2 |     93.8 |      94.5 |
+| Large       | SPATIAL_OBJ     |        95.3 |     95.5 |      95.4 |
+| Transformer | SPATIAL_OBJ     |        96.3 |     92.8 |      94.5 |

annotations.toml

@@ -0,0 +1,44 @@
+[BUILDING]
+Train = 8162
+Dev = 1811
+Test = 1540
+
+[COUNTRY]
+Train = 4227
+Dev = 849
+Test = 893
+
+[DLF]
+Train = 4305
+Dev = 998
+Test = 888
+
+[ENV_FEATURES]
+Train = 1139
+Dev = 225
+Test = 191
+
+[INT_SPACE]
+Train = 2309
+Dev = 550
+Test = 403
+
+[NPIP]
+Train = 1799
+Dev = 422
+Test = 352
+
+[POPULATED_PLACE]
+Train = 10505
+Dev = 2427
+Test = 2290
+
+[REGION]
+Train = 969
+Dev = 249
+Test = 190
+
+[SPATIAL_OBJ]
+Train = 3992
+Dev = 908
+Test = 780

assets/.gitattributes

@@ -0,0 +1,6 @@
+# This is needed to ensure that text-based assets included with project
+# templates and cloned via Git end up with consistent line endings and
+# the same checksums. It will prevent Git from converting line endings.
+# Otherwise, a user cloning assets on Windows may end up with a different
+# checksum due to different line endings.
+*	-text

assets/annotated_data.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:655a46331231574542658bf468c3716b2c76459ba7ee7d87ebc7239a16d4c535
+size 125899652

assets/annotated_data_spans.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea9e834cb3a87b06915c12e36759ff9cbeb48dbd3abdebc0c559b754105e4940
+size 112457147

assets/dev.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ae88de1d832df838613b975381734ae8a744ef08de0bde105f14acdfe2671ca8
+size 18859471

assets/test.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d42e779b0f1bad2aaed924a75777bb2280e6d58047f898d66afaa5fd7549d21b
+size 15674151

assets/train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:908cb3932321710a3f5be506eca92de804c08d63254d8f269c6ff77c33112f32
+size 77923525

configs/base_config_lg.cfg

@@ -0,0 +1,110 @@
+# This is an auto-generated partial config. To use it with 'spacy train'
+# you can run spacy init fill-config to auto-fill all default settings:
+# python -m spacy init fill-config ./base_config.cfg ./config.cfg
+[paths]
+train = null
+dev = null
+vectors = "en_core_web_lg"
+[system]
+gpu_allocator = null
+
+[nlp]
+lang = "en"
+pipeline = ["tok2vec","spancat"]
+batch_size = 1000
+
+[components]
+
+[components.tok2vec]
+factory = "tok2vec"
+
+[components.tok2vec.model]
+@architectures = "spacy.Tok2Vec.v2"
+
+[components.tok2vec.model.embed]
+@architectures = "spacy.MultiHashEmbed.v2"
+width = ${components.tok2vec.model.encode.width}
+attrs = ["NORM", "PREFIX", "SUFFIX", "SHAPE"]
+rows = [5000, 1000, 2500, 2500]
+include_static_vectors = true
+
+[components.tok2vec.model.encode]
+@architectures = "spacy.MaxoutWindowEncoder.v2"
+width = 256
+depth = 8
+window_size = 1
+maxout_pieces = 3
+
+[components.spancat]
+factory = "spancat"
+max_positive = null
+scorer = {"@scorers":"spacy.spancat_scorer.v1"}
+spans_key = "sc"
+threshold = 0.5
+
+[components.spancat.model]
+@architectures = "spacy.SpanCategorizer.v1"
+
+[components.spancat.model.reducer]
+@layers = "spacy.mean_max_reducer.v1"
+hidden_size = 128
+
+[components.spancat.model.scorer]
+@layers = "spacy.LinearLogistic.v1"
+nO = null
+nI = null
+
+[components.spancat.model.tok2vec]
+@architectures = "spacy.Tok2VecListener.v1"
+width = ${components.tok2vec.model.encode.width}
+
+[components.spancat.suggester]
+@misc = "spacy.ngram_suggester.v1"
+sizes = [1,2,3]
+
+[corpora]
+
+[corpora.train]
+@readers = "spacy.Corpus.v1"
+path = ${paths.train}
+max_length = 0
+
+[corpora.dev]
+@readers = "spacy.Corpus.v1"
+path = ${paths.dev}
+max_length = 0
+
+[training]
+dev_corpus = "corpora.dev"
+train_corpus = "corpora.train"
+seed = ${system.seed}
+gpu_allocator = ${system.gpu_allocator}
+dropout = 0.1
+accumulate_gradient = 1
+patience = 20000
+max_epochs = 10
+max_steps = 0
+eval_frequency = 200
+frozen_components = []
+annotating_components = []
+before_to_disk = null
+before_update = null
+
+[training.batcher]
+@batchers = "spacy.batch_by_words.v1"
+discard_oversize = false
+tolerance = 0.2
+get_length = null
+
+[training.batcher.size]
+@schedules = "compounding.v1"
+start = 1000
+stop = 10000
+compound = 1.001
+t = 0.0
+
+[training.optimizer]
+@optimizers = "Adam.v1"
+
+[initialize]
+vectors = ${paths.vectors}

configs/base_config_md.cfg

@@ -0,0 +1,110 @@
+# This is an auto-generated partial config. To use it with 'spacy train'
+# you can run spacy init fill-config to auto-fill all default settings:
+# python -m spacy init fill-config ./base_config.cfg ./config.cfg
+[paths]
+train = null
+dev = null
+vectors = "en_core_web_md"
+[system]
+gpu_allocator = null
+
+[nlp]
+lang = "en"
+pipeline = ["tok2vec","spancat"]
+batch_size = 1000
+
+[components]
+
+[components.tok2vec]
+factory = "tok2vec"
+
+[components.tok2vec.model]
+@architectures = "spacy.Tok2Vec.v2"
+
+[components.tok2vec.model.embed]
+@architectures = "spacy.MultiHashEmbed.v2"
+width = ${components.tok2vec.model.encode.width}
+attrs = ["NORM", "PREFIX", "SUFFIX", "SHAPE"]
+rows = [5000, 1000, 2500, 2500]
+include_static_vectors = true
+
+[components.tok2vec.model.encode]
+@architectures = "spacy.MaxoutWindowEncoder.v2"
+width = 256
+depth = 8
+window_size = 1
+maxout_pieces = 3
+
+[components.spancat]
+factory = "spancat"
+max_positive = null
+scorer = {"@scorers":"spacy.spancat_scorer.v1"}
+spans_key = "sc"
+threshold = 0.5
+
+[components.spancat.model]
+@architectures = "spacy.SpanCategorizer.v1"
+
+[components.spancat.model.reducer]
+@layers = "spacy.mean_max_reducer.v1"
+hidden_size = 128
+
+[components.spancat.model.scorer]
+@layers = "spacy.LinearLogistic.v1"
+nO = null
+nI = null
+
+[components.spancat.model.tok2vec]
+@architectures = "spacy.Tok2VecListener.v1"
+width = ${components.tok2vec.model.encode.width}
+
+[components.spancat.suggester]
+@misc = "spacy.ngram_suggester.v1"
+sizes = [1,2,3]
+
+[corpora]
+
+[corpora.train]
+@readers = "spacy.Corpus.v1"
+path = ${paths.train}
+max_length = 0
+
+[corpora.dev]
+@readers = "spacy.Corpus.v1"
+path = ${paths.dev}
+max_length = 0
+
+[training]
+dev_corpus = "corpora.dev"
+train_corpus = "corpora.train"
+seed = ${system.seed}
+gpu_allocator = ${system.gpu_allocator}
+dropout = 0.1
+accumulate_gradient = 1
+patience = 20000
+max_epochs = 10
+max_steps = 0
+eval_frequency = 200
+frozen_components = []
+annotating_components = []
+before_to_disk = null
+before_update = null
+
+[training.batcher]
+@batchers = "spacy.batch_by_words.v1"
+discard_oversize = false
+tolerance = 0.2
+get_length = null
+
+[training.batcher.size]
+@schedules = "compounding.v1"
+start = 1000
+stop = 10000
+compound = 1.001
+t = 0.0
+
+[training.optimizer]
+@optimizers = "Adam.v1"
+
+[initialize]
+vectors = ${paths.vectors}

configs/base_config_sm.cfg

@@ -0,0 +1,111 @@
+# This is an auto-generated partial config. To use it with 'spacy train'
+# you can run spacy init fill-config to auto-fill all default settings:
+# python -m spacy init fill-config ./base_config.cfg ./config.cfg
+[paths]
+train = null
+dev = null
+vectors = null
+[system]
+gpu_allocator = null
+
+[nlp]
+lang = "en"
+pipeline = ["tok2vec","spancat"]
+batch_size = 10000
+
+[components]
+
+[components.tok2vec]
+factory = "tok2vec"
+
+[components.tok2vec.model]
+@architectures = "spacy.Tok2Vec.v2"
+
+[components.tok2vec.model.embed]
+@architectures = "spacy.MultiHashEmbed.v2"
+width = ${components.tok2vec.model.encode.width}
+attrs = ["NORM", "PREFIX", "SUFFIX", "SHAPE"]
+rows = [5000, 1000, 2500, 2500]
+include_static_vectors = false
+
+[components.tok2vec.model.encode]
+@architectures = "spacy.MaxoutWindowEncoder.v2"
+width = 96
+depth = 4
+window_size = 1
+maxout_pieces = 3
+
+[components.spancat]
+factory = "spancat"
+max_positive = null
+scorer = {"@scorers":"spacy.spancat_scorer.v1"}
+spans_key = "sc"
+threshold = 0.5
+
+[components.spancat.model]
+@architectures = "spacy.SpanCategorizer.v1"
+
+[components.spancat.model.reducer]
+@layers = "spacy.mean_max_reducer.v1"
+hidden_size = 128
+
+[components.spancat.model.scorer]
+@layers = "spacy.LinearLogistic.v1"
+nO = null
+nI = null
+
+[components.spancat.model.tok2vec]
+@architectures = "spacy.Tok2VecListener.v1"
+width = ${components.tok2vec.model.encode.width}
+
+[components.spancat.suggester]
+@misc = "spacy.ngram_suggester.v1"
+sizes = [1,2,3]
+
+[corpora]
+
+[corpora.train]
+@readers = "spacy.Corpus.v1"
+path = ${paths.train}
+max_length = 0
+
+[corpora.dev]
+@readers = "spacy.Corpus.v1"
+path = ${paths.dev}
+max_length = 0
+
+[training]
+dev_corpus = "corpora.dev"
+train_corpus = "corpora.train"
+seed = ${system.seed}
+gpu_allocator = ${system.gpu_allocator}
+dropout = 0.1
+accumulate_gradient = 1
+patience = 20000
+max_epochs = 0
+max_steps = 0
+eval_frequency = 300
+frozen_components = []
+annotating_components = []
+before_to_disk = null
+before_update = null
+
+[training.checkpoints]
+save_n_steps = 1000  # Save the model every 1000 steps
+
+[training.optimizer]
+@optimizers = "Adam.v1"
+
+[training.batcher]
+@batchers = "spacy.batch_by_words.v1"
+discard_oversize = false
+tolerance = 0.2
+
+[training.batcher.size]
+@schedules = "compounding.v1"
+start = 100
+stop = 1000
+compound = 1.001
+
+[initialize]
+vectors = ${paths.vectors}

configs/base_config_trf.cfg

@@ -0,0 +1,105 @@
+# This is an auto-generated partial config. To use it with 'spacy train'
+# you can run spacy init fill-config to auto-fill all default settings:
+# python -m spacy init fill-config ./base_config.cfg ./config.cfg
+[paths]
+train = null
+dev = null
+vectors = null
+[system]
+gpu_allocator = "pytorch"
+
+[nlp]
+lang = "en"
+pipeline = ["transformer","spancat"]
+batch_size = 128
+
+[components]
+
+[components.transformer]
+factory = "transformer"
+
+[components.transformer.model]
+@architectures = "spacy-transformers.TransformerModel.v3"
+name = "roberta-base"
+tokenizer_config = {"use_fast": true}
+
+[components.transformer.model.get_spans]
+@span_getters = "spacy-transformers.strided_spans.v1"
+window = 128
+stride = 96
+
+[components.spancat]
+factory = "spancat"
+max_positive = null
+scorer = {"@scorers":"spacy.spancat_scorer.v1"}
+spans_key = "sc"
+threshold = 0.5
+
+[components.spancat.model]
+@architectures = "spacy.SpanCategorizer.v1"
+
+[components.spancat.model.reducer]
+@layers = "spacy.mean_max_reducer.v1"
+hidden_size = 128
+
+[components.spancat.model.scorer]
+@layers = "spacy.LinearLogistic.v1"
+nO = null
+nI = null
+
+[components.spancat.model.tok2vec]
+@architectures = "spacy-transformers.TransformerListener.v1"
+grad_factor = 1.0
+
+[components.spancat.model.tok2vec.pooling]
+@layers = "reduce_mean.v1"
+
+[components.spancat.suggester]
+@misc = "spacy.ngram_suggester.v1"
+sizes = [1,2,3]
+
+[corpora]
+
+[corpora.train]
+@readers = "spacy.Corpus.v1"
+path = ${paths.train}
+max_length = 0
+
+[corpora.dev]
+@readers = "spacy.Corpus.v1"
+path = ${paths.dev}
+max_length = 0
+
+[training]
+dev_corpus = "corpora.dev"
+train_corpus = "corpora.train"
+seed = ${system.seed}
+gpu_allocator = ${system.gpu_allocator}
+dropout = 0.1
+accumulate_gradient = 1
+patience = 20000
+max_epochs = 10
+max_steps = 0
+eval_frequency = 200
+frozen_components = []
+annotating_components = []
+before_to_disk = null
+before_update = null
+
+[training.optimizer]
+@optimizers = "Adam.v1"
+
+[training.optimizer.learn_rate]
+@schedules = "warmup_linear.v1"
+warmup_steps = 250
+total_steps = 20000
+initial_rate = 5e-5
+
+[training.batcher]
+@batchers = "spacy.batch_by_padded.v1"
+discard_oversize = true
+size = 2000
+buffer = 256
+
+[initialize]
+vectors = ${paths.vectors}

configs/config.cfg

@@ -0,0 +1,151 @@
+[paths]
+train = null
+dev = null
+vectors = null
+init_tok2vec = null
+
+[system]
+gpu_allocator = null
+seed = 0
+
+[nlp]
+lang = "en"
+pipeline = ["tok2vec","spancat"]
+batch_size = 10000
+disabled = []
+before_creation = null
+after_creation = null
+after_pipeline_creation = null
+tokenizer = {"@tokenizers":"spacy.Tokenizer.v1"}
+vectors = {"@vectors":"spacy.Vectors.v1"}
+
+[components]
+
+[components.spancat]
+factory = "spancat"
+max_positive = null
+scorer = {"@scorers":"spacy.spancat_scorer.v1"}
+spans_key = "sc"
+threshold = 0.5
+
+[components.spancat.model]
+@architectures = "spacy.SpanCategorizer.v1"
+
+[components.spancat.model.reducer]
+@layers = "spacy.mean_max_reducer.v1"
+hidden_size = 128
+
+[components.spancat.model.scorer]
+@layers = "spacy.LinearLogistic.v1"
+nO = null
+nI = null
+
+[components.spancat.model.tok2vec]
+@architectures = "spacy.Tok2VecListener.v1"
+width = ${components.tok2vec.model.encode.width}
+upstream = "*"
+
+[components.spancat.suggester]
+@misc = "spacy.ngram_suggester.v1"
+sizes = [1,2,3]
+
+[components.tok2vec]
+factory = "tok2vec"
+
+[components.tok2vec.model]
+@architectures = "spacy.Tok2Vec.v2"
+
+[components.tok2vec.model.embed]
+@architectures = "spacy.MultiHashEmbed.v2"
+width = ${components.tok2vec.model.encode.width}
+attrs = ["NORM","PREFIX","SUFFIX","SHAPE"]
+rows = [5000,1000,2500,2500]
+include_static_vectors = false
+
+[components.tok2vec.model.encode]
+@architectures = "spacy.MaxoutWindowEncoder.v2"
+width = 96
+depth = 4
+window_size = 1
+maxout_pieces = 3
+
+[corpora]
+
+[corpora.dev]
+@readers = "spacy.Corpus.v1"
+path = ${paths.dev}
+max_length = 0
+gold_preproc = false
+limit = 0
+augmenter = null
+
+[corpora.train]
+@readers = "spacy.Corpus.v1"
+path = ${paths.train}
+max_length = 0
+gold_preproc = false
+limit = 0
+augmenter = null
+
+[training]
+dev_corpus = "corpora.dev"
+train_corpus = "corpora.train"
+seed = ${system.seed}
+gpu_allocator = ${system.gpu_allocator}
+dropout = 0.1
+accumulate_gradient = 1
+patience = 20000
+max_epochs = 10
+max_steps = 0
+eval_frequency = 200
+frozen_components = []
+annotating_components = []
+before_to_disk = null
+before_update = null
+
+[training.batcher]
+@batchers = "spacy.batch_by_words.v1"
+discard_oversize = false
+tolerance = 0.2
+get_length = null
+
+[training.batcher.size]
+@schedules = "compounding.v1"
+start = 1000
+stop = 10000
+compound = 1.001
+t = 0.0
+
+[training.logger]
+@loggers = "spacy.ConsoleLogger.v1"
+progress_bar = false
+
+[training.optimizer]
+@optimizers = "Adam.v1"
+beta1 = 0.9
+beta2 = 0.999
+L2_is_weight_decay = true
+L2 = 0.01
+grad_clip = 1.0
+use_averages = false
+eps = 0.00000001
+learn_rate = 0.001
+
+[training.score_weights]
+spans_sc_f = 1.0
+spans_sc_p = 0.0
+spans_sc_r = 0.0
+
+[pretraining]
+
+[initialize]
+vectors = ${paths.vectors}
+init_tok2vec = ${paths.init_tok2vec}
+vocab_data = null
+lookups = null
+before_init = null
+after_init = null
+
+[initialize.components]
+
+[initialize.tokenizer]

configs/config_lg.cfg

@@ -0,0 +1,151 @@
+[paths]
+train = null
+dev = null
+vectors = "en_core_web_lg"
+init_tok2vec = null
+
+[system]
+gpu_allocator = null
+seed = 0
+
+[nlp]
+lang = "en"
+pipeline = ["tok2vec","spancat"]
+batch_size = 1000
+disabled = []
+before_creation = null
+after_creation = null
+after_pipeline_creation = null
+tokenizer = {"@tokenizers":"spacy.Tokenizer.v1"}
+vectors = {"@vectors":"spacy.Vectors.v1"}
+
+[components]
+
+[components.spancat]
+factory = "spancat"
+max_positive = null
+scorer = {"@scorers":"spacy.spancat_scorer.v1"}
+spans_key = "sc"
+threshold = 0.5
+
+[components.spancat.model]
+@architectures = "spacy.SpanCategorizer.v1"
+
+[components.spancat.model.reducer]
+@layers = "spacy.mean_max_reducer.v1"
+hidden_size = 128
+
+[components.spancat.model.scorer]
+@layers = "spacy.LinearLogistic.v1"
+nO = null
+nI = null
+
+[components.spancat.model.tok2vec]
+@architectures = "spacy.Tok2VecListener.v1"
+width = ${components.tok2vec.model.encode.width}
+upstream = "*"
+
+[components.spancat.suggester]
+@misc = "spacy.ngram_suggester.v1"
+sizes = [1,2,3]
+
+[components.tok2vec]
+factory = "tok2vec"
+
+[components.tok2vec.model]
+@architectures = "spacy.Tok2Vec.v2"
+
+[components.tok2vec.model.embed]
+@architectures = "spacy.MultiHashEmbed.v2"
+width = ${components.tok2vec.model.encode.width}
+attrs = ["NORM","PREFIX","SUFFIX","SHAPE"]
+rows = [5000,1000,2500,2500]
+include_static_vectors = true
+
+[components.tok2vec.model.encode]
+@architectures = "spacy.MaxoutWindowEncoder.v2"
+width = 256
+depth = 8
+window_size = 1
+maxout_pieces = 3
+
+[corpora]
+
+[corpora.dev]
+@readers = "spacy.Corpus.v1"
+path = ${paths.dev}
+max_length = 0
+gold_preproc = false
+limit = 0
+augmenter = null
+
+[corpora.train]
+@readers = "spacy.Corpus.v1"
+path = ${paths.train}
+max_length = 0
+gold_preproc = false
+limit = 0
+augmenter = null
+
+[training]
+dev_corpus = "corpora.dev"
+train_corpus = "corpora.train"
+seed = ${system.seed}
+gpu_allocator = ${system.gpu_allocator}
+dropout = 0.1
+accumulate_gradient = 1
+patience = 20000
+max_epochs = 10
+max_steps = 0
+eval_frequency = 200
+frozen_components = []
+annotating_components = []
+before_to_disk = null
+before_update = null
+
+[training.batcher]
+@batchers = "spacy.batch_by_words.v1"
+discard_oversize = false
+tolerance = 0.2
+get_length = null
+
+[training.batcher.size]
+@schedules = "compounding.v1"
+start = 1000
+stop = 10000
+compound = 1.001
+t = 0.0
+
+[training.logger]
+@loggers = "spacy.ConsoleLogger.v1"
+progress_bar = false
+
+[training.optimizer]
+@optimizers = "Adam.v1"
+beta1 = 0.9
+beta2 = 0.999
+L2_is_weight_decay = true
+L2 = 0.01
+grad_clip = 1.0
+use_averages = false
+eps = 0.00000001
+learn_rate = 0.001
+
+[training.score_weights]
+spans_sc_f = 1.0
+spans_sc_p = 0.0
+spans_sc_r = 0.0
+
+[pretraining]
+
+[initialize]
+vectors = ${paths.vectors}
+init_tok2vec = ${paths.init_tok2vec}
+vocab_data = null
+lookups = null
+before_init = null
+after_init = null
+
+[initialize.components]
+
+[initialize.tokenizer]

configs/config_md.cfg

@@ -0,0 +1,151 @@
+[paths]
+train = null
+dev = null
+vectors = "en_core_web_md"
+init_tok2vec = null
+
+[system]
+gpu_allocator = null
+seed = 0
+
+[nlp]
+lang = "en"
+pipeline = ["tok2vec","spancat"]
+batch_size = 1000
+disabled = []
+before_creation = null
+after_creation = null
+after_pipeline_creation = null
+tokenizer = {"@tokenizers":"spacy.Tokenizer.v1"}
+vectors = {"@vectors":"spacy.Vectors.v1"}
+
+[components]
+
+[components.spancat]
+factory = "spancat"
+max_positive = null
+scorer = {"@scorers":"spacy.spancat_scorer.v1"}
+spans_key = "sc"
+threshold = 0.5
+
+[components.spancat.model]
+@architectures = "spacy.SpanCategorizer.v1"
+
+[components.spancat.model.reducer]
+@layers = "spacy.mean_max_reducer.v1"
+hidden_size = 128
+
+[components.spancat.model.scorer]
+@layers = "spacy.LinearLogistic.v1"
+nO = null
+nI = null
+
+[components.spancat.model.tok2vec]
+@architectures = "spacy.Tok2VecListener.v1"
+width = ${components.tok2vec.model.encode.width}
+upstream = "*"
+
+[components.spancat.suggester]
+@misc = "spacy.ngram_suggester.v1"
+sizes = [1,2,3]
+
+[components.tok2vec]
+factory = "tok2vec"
+
+[components.tok2vec.model]
+@architectures = "spacy.Tok2Vec.v2"
+
+[components.tok2vec.model.embed]
+@architectures = "spacy.MultiHashEmbed.v2"
+width = ${components.tok2vec.model.encode.width}
+attrs = ["NORM","PREFIX","SUFFIX","SHAPE"]
+rows = [5000,1000,2500,2500]
+include_static_vectors = true
+
+[components.tok2vec.model.encode]
+@architectures = "spacy.MaxoutWindowEncoder.v2"
+width = 256
+depth = 8
+window_size = 1
+maxout_pieces = 3
+
+[corpora]
+
+[corpora.dev]
+@readers = "spacy.Corpus.v1"
+path = ${paths.dev}
+max_length = 0
+gold_preproc = false
+limit = 0
+augmenter = null
+
+[corpora.train]
+@readers = "spacy.Corpus.v1"
+path = ${paths.train}
+max_length = 0
+gold_preproc = false
+limit = 0
+augmenter = null
+
+[training]
+dev_corpus = "corpora.dev"
+train_corpus = "corpora.train"
+seed = ${system.seed}
+gpu_allocator = ${system.gpu_allocator}
+dropout = 0.1
+accumulate_gradient = 1
+patience = 20000
+max_epochs = 10
+max_steps = 0
+eval_frequency = 200
+frozen_components = []
+annotating_components = []
+before_to_disk = null
+before_update = null
+
+[training.batcher]
+@batchers = "spacy.batch_by_words.v1"
+discard_oversize = false
+tolerance = 0.2
+get_length = null
+
+[training.batcher.size]
+@schedules = "compounding.v1"
+start = 1000
+stop = 10000
+compound = 1.001
+t = 0.0
+
+[training.logger]
+@loggers = "spacy.ConsoleLogger.v1"
+progress_bar = false
+
+[training.optimizer]
+@optimizers = "Adam.v1"
+beta1 = 0.9
+beta2 = 0.999
+L2_is_weight_decay = true
+L2 = 0.01
+grad_clip = 1.0
+use_averages = false
+eps = 0.00000001
+learn_rate = 0.001
+
+[training.score_weights]
+spans_sc_f = 1.0
+spans_sc_p = 0.0
+spans_sc_r = 0.0
+
+[pretraining]
+
+[initialize]
+vectors = ${paths.vectors}
+init_tok2vec = ${paths.init_tok2vec}
+vocab_data = null
+lookups = null
+before_init = null
+after_init = null
+
+[initialize.components]
+
+[initialize.tokenizer]

configs/config_sm.cfg

@@ -0,0 +1,151 @@
+[paths]
+train = null
+dev = null
+vectors = null
+init_tok2vec = null
+
+[system]
+gpu_allocator = null
+seed = 0
+
+[nlp]
+lang = "en"
+pipeline = ["tok2vec","spancat"]
+batch_size = 1000
+disabled = []
+before_creation = null
+after_creation = null
+after_pipeline_creation = null
+tokenizer = {"@tokenizers":"spacy.Tokenizer.v1"}
+vectors = {"@vectors":"spacy.Vectors.v1"}
+
+[components]
+
+[components.spancat]
+factory = "spancat"
+max_positive = null
+scorer = {"@scorers":"spacy.spancat_scorer.v1"}
+spans_key = "sc"
+threshold = 0.5
+
+[components.spancat.model]
+@architectures = "spacy.SpanCategorizer.v1"
+
+[components.spancat.model.reducer]
+@layers = "spacy.mean_max_reducer.v1"
+hidden_size = 128
+
+[components.spancat.model.scorer]
+@layers = "spacy.LinearLogistic.v1"
+nO = null
+nI = null
+
+[components.spancat.model.tok2vec]
+@architectures = "spacy.Tok2VecListener.v1"
+width = ${components.tok2vec.model.encode.width}
+upstream = "*"
+
+[components.spancat.suggester]
+@misc = "spacy.ngram_suggester.v1"
+sizes = [1,2,3]
+
+[components.tok2vec]
+factory = "tok2vec"
+
+[components.tok2vec.model]
+@architectures = "spacy.Tok2Vec.v2"
+
+[components.tok2vec.model.embed]
+@architectures = "spacy.MultiHashEmbed.v2"
+width = ${components.tok2vec.model.encode.width}
+attrs = ["NORM","PREFIX","SUFFIX","SHAPE"]
+rows = [5000,1000,2500,2500]
+include_static_vectors = false
+
+[components.tok2vec.model.encode]
+@architectures = "spacy.MaxoutWindowEncoder.v2"
+width = 96
+depth = 4
+window_size = 1
+maxout_pieces = 3
+
+[corpora]
+
+[corpora.dev]
+@readers = "spacy.Corpus.v1"
+path = ${paths.dev}
+max_length = 0
+gold_preproc = false
+limit = 0
+augmenter = null
+
+[corpora.train]
+@readers = "spacy.Corpus.v1"
+path = ${paths.train}
+max_length = 0
+gold_preproc = false
+limit = 0
+augmenter = null
+
+[training]
+dev_corpus = "corpora.dev"
+train_corpus = "corpora.train"
+seed = ${system.seed}
+gpu_allocator = ${system.gpu_allocator}
+dropout = 0.1
+accumulate_gradient = 1
+patience = 200
+max_epochs = 10
+max_steps = 0
+eval_frequency = 20
+frozen_components = []
+annotating_components = []
+before_to_disk = null
+before_update = null
+
+[training.batcher]
+@batchers = "spacy.batch_by_words.v1"
+discard_oversize = false
+tolerance = 0.2
+get_length = null
+
+[training.batcher.size]
+@schedules = "compounding.v1"
+start = 1000
+stop = 10000
+compound = 1.001
+t = 0.0
+
+[training.logger]
+@loggers = "spacy.ConsoleLogger.v1"
+progress_bar = false
+
+[training.optimizer]
+@optimizers = "Adam.v1"
+beta1 = 0.9
+beta2 = 0.999
+L2_is_weight_decay = true
+L2 = 0.01
+grad_clip = 1.0
+use_averages = false
+eps = 0.00000001
+learn_rate = 0.001
+
+[training.score_weights]
+spans_sc_f = 1.0
+spans_sc_p = 0.0
+spans_sc_r = 0.0
+
+[pretraining]
+
+[initialize]
+vectors = ${paths.vectors}
+init_tok2vec = ${paths.init_tok2vec}
+vocab_data = null
+lookups = null
+before_init = null
+after_init = null
+
+[initialize.components]
+
+[initialize.tokenizer]

configs/config_trf.cfg

@@ -0,0 +1,153 @@
+[paths]
+train = null
+dev = null
+vectors = null
+init_tok2vec = null
+
+[system]
+gpu_allocator = "pytorch"
+seed = 0
+
+[nlp]
+lang = "en"
+pipeline = ["transformer","spancat"]
+batch_size = 5000
+disabled = []
+before_creation = null
+after_creation = null
+after_pipeline_creation = null
+tokenizer = {"@tokenizers":"spacy.Tokenizer.v1"}
+vectors = {"@vectors":"spacy.Vectors.v1"}
+
+[components]
+
+[components.spancat]
+factory = "spancat"
+max_positive = null
+scorer = {"@scorers":"spacy.spancat_scorer.v1"}
+spans_key = "sc"
+threshold = 0.5
+
+[components.spancat.model]
+@architectures = "spacy.SpanCategorizer.v1"
+
+[components.spancat.model.reducer]
+@layers = "spacy.mean_max_reducer.v1"
+hidden_size = 128
+
+[components.spancat.model.scorer]
+@layers = "spacy.LinearLogistic.v1"
+nO = null
+nI = null
+
+[components.spancat.model.tok2vec]
+@architectures = "spacy-transformers.TransformerListener.v1"
+grad_factor = 1.0
+pooling = {"@layers":"reduce_mean.v1"}
+upstream = "*"
+
+[components.spancat.suggester]
+@misc = "spacy.ngram_suggester.v1"
+sizes = [1,2,3]
+
+[components.transformer]
+factory = "transformer"
+max_batch_items = 4096
+set_extra_annotations = {"@annotation_setters":"spacy-transformers.null_annotation_setter.v1"}
+
+[components.transformer.model]
+@architectures = "spacy-transformers.TransformerModel.v3"
+name = "roberta-base"
+mixed_precision = false
+
+[components.transformer.model.get_spans]
+@span_getters = "spacy-transformers.strided_spans.v1"
+window = 128
+stride = 96
+
+[components.transformer.model.grad_scaler_config]
+
+[components.transformer.model.tokenizer_config]
+use_fast = true
+
+[components.transformer.model.transformer_config]
+
+[corpora]
+
+[corpora.dev]
+@readers = "spacy.Corpus.v1"
+path = ${paths.dev}
+max_length = 0
+gold_preproc = false
+limit = 0
+augmenter = null
+
+[corpora.train]
+@readers = "spacy.Corpus.v1"
+path = ${paths.train}
+max_length = 0
+gold_preproc = false
+limit = 0
+augmenter = null
+
+[training]
+dev_corpus = "corpora.dev"
+train_corpus = "corpora.train"
+seed = ${system.seed}
+gpu_allocator = ${system.gpu_allocator}
+dropout = 0.1
+accumulate_gradient = 1
+patience = 20000
+max_epochs = 10
+max_steps = 0
+eval_frequency = 200
+frozen_components = []
+annotating_components = []
+before_to_disk = null
+before_update = null
+
+[training.batcher]
+@batchers = "spacy.batch_by_padded.v1"
+discard_oversize = true
+size = 2000
+buffer = 256
+get_length = null
+
+[training.logger]
+@loggers = "spacy.ConsoleLogger.v1"
+progress_bar = false
+
+[training.optimizer]
+@optimizers = "Adam.v1"
+beta1 = 0.9
+beta2 = 0.999
+L2_is_weight_decay = true
+L2 = 0.01
+grad_clip = 1.0
+use_averages = false
+eps = 0.00000001
+
+[training.optimizer.learn_rate]
+@schedules = "warmup_linear.v1"
+warmup_steps = 250
+total_steps = 20000
+initial_rate = 0.00005
+
+[training.score_weights]
+spans_sc_f = 1.0
+spans_sc_p = 0.0
+spans_sc_r = 0.0
+
+[pretraining]
+
+[initialize]
+vectors = ${paths.vectors}
+init_tok2vec = ${paths.init_tok2vec}
+vocab_data = null
+lookups = null
+before_init = null
+after_init = null
+
+[initialize.components]
+
+[initialize.tokenizer]

corpus/dev.spacy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1560591d4e8a0064c6c359a322d89b679935c6cc530f5fdcfbe553b8e06f21c3
+size 918183

corpus/test.spacy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fe99e1460aa508a681da92b37c134841509de398933a6a990eb479dd1de87c94
+size 787612

corpus/train.spacy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d06e421c115d51d076d660262478e9cf5ca18b01c249b4bf93c26d1be3aeec30
+size 3869194

gold-training-data/christine_0020_0040_annotated.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9866f542e3e0a52ed20dc101448bd9a03254e714bf8e8ba6b164523356e06303
+size 31070680

gold-training-data/christine_0040_0060.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:edd79c66e3b4fbe5eeb4492e0aa1259e5a3ee879df75e2bac707be7056cb8537
+size 49669585

gold-training-data/greg_0000_0020.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84a946d422975a14022c16efdb8d7f05d5ef60c935c47e7eb57ca97ad7292bb7
+size 22953073

gold-training-data/greg_60_80.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ed22a3c20a95fdd2afd900350bae6e0f46a93dc698ea60f21da29b79d92c67aa
+size 22206314

model_comparison.md

@@ -0,0 +1,47 @@
+# Overall Model Performance
+| Model       |   Precision |   Recall |   F-Score |
+|:------------|------------:|---------:|----------:|
+| Small       |        94.1 |     89.2 |      91.6 |
+| Medium      |        94   |     90.5 |      92.2 |
+| Large       |        94.1 |     91.7 |      92.9 |
+| Transformer |        93.6 |     91.6 |      92.6 |
+
+# Performance per Label
+| Model       | Label           |   Precision |   Recall |   F-Score |
+|:------------|:----------------|------------:|---------:|----------:|
+| Small       | BUILDING        |        94.7 |     90.2 |      92.4 |
+| Medium      | BUILDING        |        95.2 |     92.8 |      94   |
+| Large       | BUILDING        |        94.8 |     93.2 |      94   |
+| Transformer | BUILDING        |        94.3 |     94.2 |      94.3 |
+| Small       | COUNTRY         |        97.6 |     94.6 |      96.1 |
+| Medium      | COUNTRY         |        96.5 |     96.3 |      96.4 |
+| Large       | COUNTRY         |        97.7 |     96.8 |      97.2 |
+| Transformer | COUNTRY         |        96.6 |     96.8 |      96.7 |
+| Small       | DLF             |        92.4 |     86.4 |      89.3 |
+| Medium      | DLF             |        95   |     84.1 |      89.2 |
+| Large       | DLF             |        93.5 |     88.4 |      90.9 |
+| Transformer | DLF             |        94.1 |     90.4 |      92.2 |
+| Small       | ENV_FEATURES    |        86.6 |     81.2 |      83.8 |
+| Medium      | ENV_FEATURES    |        86.3 |     79.1 |      82.5 |
+| Large       | ENV_FEATURES    |        77.5 |     90.1 |      83.3 |
+| Transformer | ENV_FEATURES    |        85.1 |     86.9 |      86   |
+| Small       | INT_SPACE       |        93.8 |     85.9 |      89.6 |
+| Medium      | INT_SPACE       |        93.9 |     91.3 |      92.6 |
+| Large       | INT_SPACE       |        92.4 |     93.8 |      93.1 |
+| Transformer | INT_SPACE       |        94.6 |     91.8 |      93.2 |
+| Small       | NPIP            |        92.7 |     86.4 |      89.4 |
+| Medium      | NPIP            |        94.5 |     82.4 |      88   |
+| Large       | NPIP            |        92.7 |     86.6 |      89.6 |
+| Transformer | NPIP            |        94.8 |     83   |      88.5 |
+| Small       | POPULATED_PLACE |        94   |     90.6 |      92.3 |
+| Medium      | POPULATED_PLACE |        93   |     91.2 |      92.1 |
+| Large       | POPULATED_PLACE |        95.2 |     90.4 |      92.7 |
+| Transformer | POPULATED_PLACE |        92.1 |     91.3 |      91.7 |
+| Small       | REGION          |        84.4 |     68.4 |      75.6 |
+| Medium      | REGION          |        81.4 |     75.8 |      78.5 |
+| Large       | REGION          |        83   |     76.8 |      79.8 |
+| Transformer | REGION          |        81.2 |     68.4 |      74.3 |
+| Small       | SPATIAL_OBJ     |        96   |     90   |      92.9 |
+| Medium      | SPATIAL_OBJ     |        95.2 |     93.8 |      94.5 |
+| Large       | SPATIAL_OBJ     |        95.3 |     95.5 |      95.4 |
+| Transformer | SPATIAL_OBJ     |        96.3 |     92.8 |      94.5 |

model_comparison.tex

@@ -0,0 +1,17 @@
+\begin{tabular}{l|cccc|cccc|cccc|cccc}
+\toprule
+{} & \multicolumn{4}{c}{Precision} & \multicolumn{4}{c}{Recall} & \multicolumn{4}{c}{F-Score} \\
+\textbf{Model} &     Large & Medium & Small & Transformer &  Large & Medium & Small & Transformer &   Large & Medium & Small & Transformer \\
+\textbf{Label          } &           &        &       &             &        &        &       &             &         &        &       &             \\
+\midrule
+\textbf{BUILDING       } &      94.8 &   95.2 &  94.7 &        94.3 &   93.2 &   92.8 &  90.2 &        94.2 &    94.0 &   94.0 &  92.4 &        94.3 \\
+\textbf{COUNTRY        } &      97.7 &   96.5 &  97.6 &        96.6 &   96.8 &   96.3 &  94.6 &        96.8 &    97.2 &   96.4 &  96.1 &        96.7 \\
+\textbf{DLF            } &      93.5 &   95.0 &  92.4 &        94.1 &   88.4 &   84.1 &  86.4 &        90.4 &    90.9 &   89.2 &  89.3 &        92.2 \\
+\textbf{ENV_FEATURES   } &      77.5 &   86.3 &  86.6 &        85.1 &   90.1 &   79.1 &  81.2 &        86.9 &    83.3 &   82.5 &  83.8 &        86.0 \\
+\textbf{INT_SPACE      } &      92.4 &   93.9 &  93.8 &        94.6 &   93.8 &   91.3 &  85.9 &        91.8 &    93.1 &   92.6 &  89.6 &        93.2 \\
+\textbf{NPIP           } &      92.7 &   94.5 &  92.7 &        94.8 &   86.6 &   82.4 &  86.4 &        83.0 &    89.6 &   88.0 &  89.4 &        88.5 \\
+\textbf{POPULATED_PLACE} &      95.2 &   93.0 &  94.0 &        92.1 &   90.4 &   91.2 &  90.6 &        91.3 &    92.7 &   92.1 &  92.3 &        91.7 \\
+\textbf{REGION         } &      83.0 &   81.4 &  84.4 &        81.2 &   76.8 &   75.8 &  68.4 &        68.4 &    79.8 &   78.5 &  75.6 &        74.3 \\
+\textbf{SPATIAL_OBJ    } &      95.3 &   95.2 &  96.0 &        96.3 &   95.5 &   93.8 &  90.0 &        92.8 &    95.4 &   94.5 &  92.9 &        94.5 \\
+\bottomrule
+\end{tabular}

notebooks/prepare-training.ipynb

@@ -0,0 +1,246 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import srsly\n",
+    "import glob\n",
+    "from collections import Counter"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "4"
+      ]
+     },
+     "execution_count": 10,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "files = glob.glob(\"./gold-training-data/*.jsonl\")\n",
+    "len(files)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "all_data = []\n",
+    "for filename in files:\n",
+    "    data = list(srsly.read_jsonl(filename))\n",
+    "    for item in data:\n",
+    "        if len(item[\"spans\"]) > 0:\n",
+    "            all_data.append(item)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "7868"
+      ]
+     },
+     "execution_count": 12,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "len(all_data)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'text': 'I was born in a small town called , and I was born May 5, 1928.',\n",
+       " 'spans': [{'start': 22,\n",
+       "   'end': 26,\n",
+       "   'token_start': 6,\n",
+       "   'token_end': 6,\n",
+       "   'label': 'POPULATED_PLACE'}],\n",
+       " '_input_hash': 1949719959,\n",
+       " '_task_hash': 335893137,\n",
+       " 'tokens': [{'text': 'I', 'start': 0, 'end': 1, 'id': 0, 'ws': True},\n",
+       "  {'text': 'was', 'start': 2, 'end': 5, 'id': 1, 'ws': True},\n",
+       "  {'text': 'born', 'start': 6, 'end': 10, 'id': 2, 'ws': True},\n",
+       "  {'text': 'in', 'start': 11, 'end': 13, 'id': 3, 'ws': True},\n",
+       "  {'text': 'a', 'start': 14, 'end': 15, 'id': 4, 'ws': True},\n",
+       "  {'text': 'small', 'start': 16, 'end': 21, 'id': 5, 'ws': True},\n",
+       "  {'text': 'town', 'start': 22, 'end': 26, 'id': 6, 'ws': True},\n",
+       "  {'text': 'called', 'start': 27, 'end': 33, 'id': 7, 'ws': True},\n",
+       "  {'text': ',', 'start': 34, 'end': 35, 'id': 8, 'ws': True},\n",
+       "  {'text': 'and', 'start': 36, 'end': 39, 'id': 9, 'ws': True},\n",
+       "  {'text': 'I', 'start': 40, 'end': 41, 'id': 10, 'ws': True},\n",
+       "  {'text': 'was', 'start': 42, 'end': 45, 'id': 11, 'ws': True},\n",
+       "  {'text': 'born', 'start': 46, 'end': 50, 'id': 12, 'ws': True},\n",
+       "  {'text': 'May', 'start': 51, 'end': 54, 'id': 13, 'ws': True},\n",
+       "  {'text': '5', 'start': 55, 'end': 56, 'id': 14, 'ws': False},\n",
+       "  {'text': ',', 'start': 56, 'end': 57, 'id': 15, 'ws': True},\n",
+       "  {'text': '1928', 'start': 58, 'end': 62, 'id': 16, 'ws': False},\n",
+       "  {'text': '.', 'start': 62, 'end': 63, 'id': 17, 'ws': False}],\n",
+       " '_view_id': 'spans_manual',\n",
+       " 'answer': 'accept',\n",
+       " '_timestamp': 1669136567}"
+      ]
+     },
+     "execution_count": 13,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "all_data[0]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "POPULATED_PLACE: 15222\n",
+      "BUILDING: 11513\n",
+      "COUNTRY: 5969\n",
+      "SPATIAL_OBJ: 5680\n",
+      "DLF: 6191\n",
+      "INT_SPACE: 3262\n",
+      "ENV_FEATURES: 1555\n",
+      "REGION: 1408\n",
+      "NPIP: 2573\n"
+     ]
+    }
+   ],
+   "source": [
+    "def merge_and_deduplicate_spans(all_data):\n",
+    "    # Mapping of labels to be merged\n",
+    "    label_mapping = {\n",
+    "        'INTERIOR_SPACE': 'INT_SPACE',\n",
+    "        'RIVER': 'ENV_FEATURES',\n",
+    "        'FOREST': 'ENV_FEATURES',\n",
+    "        'GHETTO': 'POPULATED_PLACE'\n",
+    "    }\n",
+    "\n",
+    "    # Process each annotation in the dataset\n",
+    "    for annotation in all_data:\n",
+    "        new_spans = []  # List to hold updated and unique spans\n",
+    "\n",
+    "        # Process each span\n",
+    "        for span in annotation['spans']:\n",
+    "            # Skip spans with the label \"CONTINENT\"\n",
+    "            if span[\"label\"] == \"CONTINENT\":\n",
+    "                continue\n",
+    "\n",
+    "            # Update label if it's in the mapping\n",
+    "            if span['label'] in label_mapping:\n",
+    "                span['label'] = label_mapping[span['label']]\n",
+    "\n",
+    "            # Check for duplicates\n",
+    "            if span not in new_spans:\n",
+    "                new_spans.append(span)\n",
+    "\n",
+    "        # Replace old spans with new_spans\n",
+    "        annotation['spans'] = new_spans\n",
+    "    return all_data\n",
+    "\n",
+    "\n",
+    "all_data = merge_and_deduplicate_spans(all_data)\n",
+    "\n",
+    "srsly.write_jsonl(\"assets/annotated_data_spans.jsonl\", all_data)\n",
+    "\n",
+    "\n",
+    "# Create a Counter object for counting labels\n",
+    "label_counter = Counter()\n",
+    "\n",
+    "# Iterate over each annotation in the dataset\n",
+    "for annotation in all_data:\n",
+    "    # Extract labels from each 'span' in the 'spans' list and add to the counter\n",
+    "    labels = [span['label'] for span in annotation['spans']]\n",
+    "    label_counter.update(labels)\n",
+    "\n",
+    "# Print out the counts\n",
+    "for label, count in label_counter.items():\n",
+    "    print(f\"{label}: {count}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "7868"
+      ]
+     },
+     "execution_count": 15,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "len(all_data)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "holocaust",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.13"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

notebooks/testing.ipynb

@@ -0,0 +1,1010 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import spacy\n",
+    "from spacy import displacy"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "text = \"\"\"\n",
+    "ANSWER: None QUESTION: Please tell us your full name.\n",
+    "\n",
+    "ANSWER: Marie Zosnika Schwartzman.\n",
+    "\n",
+    "QUESTION: And where were you born?\n",
+    "\n",
+    "ANSWER: In Paris . Day, August 21st, 1925.\n",
+    "\n",
+    "QUESTION: Let's talk about your family. What was your father's name?\n",
+    "\n",
+    "ANSWER: Avram.\n",
+    "\n",
+    "QUESTION: And what did he do?\n",
+    "\n",
+    "ANSWER: He was like a clothes designer.\n",
+    "\n",
+    "QUESTION: And where -- where did he come from?\n",
+    "\n",
+    "ANSWER: Poland .\n",
+    "\n",
+    "QUESTION: What -- what city in Poland ?\n",
+    "\n",
+    "ANSWER: Ciechanow . Don't ask me to spell it.\n",
+    "\n",
+    "QUESTION: Okay. And you -- ANSWER: But I think it's in the thing there.\n",
+    "\n",
+    "QUESTION: And what was your mother's name?\n",
+    "\n",
+    "ANSWER: Tovah.\n",
+    "\n",
+    "QUESTION: And her maiden name? Do you know that?\n",
+    "\n",
+    "ANSWER: Mifda (phonetic). Q: And where was she from originally?\n",
+    "\n",
+    "ANSWER: The same town , Ciechanow .\n",
+    "\n",
+    "QUESTION: Uh-huh. And was your family a very religious family?\n",
+    "\n",
+    "ANSWER: I don't think so. Not my father anyway. He was not religious. I think my mother was a little bit more religious. But we were not raised religiously. Just we kept to traditions and that's it.\n",
+    "\n",
+    "QUESTION: Now, you said you were born in Paris .\n",
+    "\n",
+    "ANSWER: Yes.\n",
+    "\n",
+    "QUESTION: When did your family come to Paris ?\n",
+    "\n",
+    "ANSWER: In 1922.\n",
+    "\n",
+    "QUESTION: What brought them there?\n",
+    "\n",
+    "ANSWER: My father came much earlier. And my family -- my mother joined with my oldest sister in Paris -- to Paris .\n",
+    "\n",
+    "QUESTION: What brought your father to Paris ?\n",
+    "\n",
+    "ANSWER: I think the Army, to a certain extent, and a job, a good job.\n",
+    "\n",
+    "QUESTION: How large was your family?\n",
+    "\n",
+    "ANSWER: Well, I had four sisters, two brothers, and my mother and father. I had a lot of family in Poland ; my grandparents, my aunts, my uncles, my cousins.\n",
+    "\n",
+    "QUESTION: But you were born in Paris . Did you go back to Poland to visit?\n",
+    "\n",
+    "ANSWER: Yes, my grandfather, we went to visit.\n",
+    "\n",
+    "QUESTION: And you stayed with large extended family there?\n",
+    "\n",
+    "ANSWER: Yes, my grandfather didn't make my mother go home , go back.\n",
+    "\n",
+    "QUESTION: He wanted her to stay?\n",
+    "\n",
+    "ANSWER: Yes.\n",
+    "\n",
+    "QUESTION: Yeah. And so where were you in the order of children? Were you the second?\n",
+    "\n",
+    "ANSWER: Second.\n",
+    "\n",
+    "QUESTION: You were the second child?\n",
+    "\n",
+    "ANSWER: Right.\n",
+    "\n",
+    "QUESTION: Yeah. And what kind of neighborhood did you live in in Paris ?\n",
+    "\n",
+    "ANSWER: We lived in a mixed neighborhood .\n",
+    "\n",
+    "QUESTION: Non-Jews and -- ANSWER: Non-Jews and Jews.\n",
+    "\n",
+    "QUESTION: Uh-huh. And did you have non-Jewish friends?\n",
+    "\n",
+    "ANSWER: Yes. Actually, I was -- when I came back from the camp , I lived with my friend I went to school with when we was six and a half years old.\n",
+    "\n",
+    "QUESTION: Uh-huh.\n",
+    "\n",
+    "ANSWER: And she was Catholic.\n",
+    "\n",
+    "QUESTION: Uh-huh.\n",
+    "\n",
+    "ANSWER: My Jewish family didn't have room for me.\n",
+    "\n",
+    "QUESTION: Uh-huh. Let's talk a little bit about your schooling. When did you begin school ? How old were you?\n",
+    "\n",
+    "ANSWER: We begin school early in France . We go to first the maternelle, which was maybe about three years old. And the -- QUESTION: Is this a public school ?\n",
+    "\n",
+    "ANSWER: Yes.\n",
+    "\n",
+    "QUESTION: So there were Jews and non-Jews in your class?\n",
+    "\n",
+    "ANSWER: Yes, Jews and non-Jews. But mostly there were non-Jews in my school . And then first grade, we start -- I think I was six, six and a half. And when I was arrested, I was in lycee . And didn't finish my education, therefore. Q: Yeah. Yeah.\n",
+    "\n",
+    "ANSWER: Six years.\n",
+    "\n",
+    "QUESTION: Yeah. We'll get to that in a few minutes. So you -- you went to a regular public school .\n",
+    "\n",
+    "ANSWER: Uh-huh.\n",
+    "\n",
+    "QUESTION: That's now what we call elementary school .\n",
+    "\n",
+    "ANSWER: Yes.\n",
+    "\n",
+    "QUESTION: Was there any problems there? You said 1t was mixed.\n",
+    "\n",
+    "ANSWER: Yeah. No problem.\n",
+    "\n",
+    "QUESTION: No problems?\n",
+    "\n",
+    "ANSWER: No.\n",
+    "\n",
+    "QUESTION: You were accepted?\n",
+    "\n",
+    "ANSWER: Yes. No problem. On the contrary, our teachers liked us very much. We were five girls in the same school , because we had girls' schools and boys' schools . It was not mixed schools . And the -- my principal, when we were arrested, came every single day to see if there were news from us. So it was no problem at all.\n",
+    "\n",
+    "QUESTION: Uh-huh. What kind of neighborhood did you live in? Was it upper class or middle class neighborhood?\n",
+    "\n",
+    "ANSWER: Probably middle class, I would say.\n",
+    "\n",
+    "QUESTION: Uh-huh. Did you live in a house or an apartment ?\n",
+    "\n",
+    "ANSWER: A house .\n",
+    "\n",
+    "QUESTION: And was it in the center of the city or the outskirts?\n",
+    "\"\"\""
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "nlp = spacy.load(\"training/md/model-best/\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "doc = nlp(text)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<span class=\"tex2jax_ignore\"><div class=\"spans\" style=\"line-height: 2.5; direction: ltr\">\n",
+       " ANSWER : None QUESTION : Please tell us your full name . \n",
+       "\n",
+       " ANSWER : Marie Zosnika Schwartzman . \n",
+       "\n",
+       " QUESTION : And where were you born ? \n",
+       "\n",
+       " ANSWER : In \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    Paris\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        POPULATED_PLACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       ". Day , August 21st , 1925 . \n",
+       "\n",
+       " QUESTION : Let 's talk about your family . What was your father 's name ? \n",
+       "\n",
+       " ANSWER : Avram . \n",
+       "\n",
+       " QUESTION : And what did he do ? \n",
+       "\n",
+       " ANSWER : He was like a clothes designer . \n",
+       "\n",
+       " QUESTION : And where -- where did he come from ? \n",
+       "\n",
+       " ANSWER : \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    Poland\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        COUNTRY\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       ". \n",
+       "\n",
+       " QUESTION : What -- what \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    city\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        POPULATED_PLACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "in \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    Poland\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        COUNTRY\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "? \n",
+       "\n",
+       " ANSWER : Ciechanow . Do n't ask me to spell it . \n",
+       "\n",
+       " QUESTION : Okay . And you -- ANSWER : But I think it 's in the thing there . \n",
+       "\n",
+       " QUESTION : And what was your mother 's name ? \n",
+       "\n",
+       " ANSWER : Tovah . \n",
+       "\n",
+       " QUESTION : And her maiden name ? Do you know that ? \n",
+       "\n",
+       " ANSWER : Mifda ( phonetic ) . Q : And where was she from originally ? \n",
+       "\n",
+       " ANSWER : The same \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    town\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        POPULATED_PLACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       ", \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    Ciechanow\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        POPULATED_PLACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       ". \n",
+       "\n",
+       " QUESTION : Uh - huh . And was your family a very religious family ? \n",
+       "\n",
+       " ANSWER : I do n't think so . Not my father anyway . He was not religious . I think my mother was a little bit more religious . But we were not raised religiously . Just we kept to traditions and that 's it . \n",
+       "\n",
+       " QUESTION : Now , you said you were born in \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    Paris\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        POPULATED_PLACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       ". \n",
+       "\n",
+       " ANSWER : Yes . \n",
+       "\n",
+       " QUESTION : When did your family come to \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    Paris\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        POPULATED_PLACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "? \n",
+       "\n",
+       " ANSWER : In 1922 . \n",
+       "\n",
+       " QUESTION : What brought them there ? \n",
+       "\n",
+       " ANSWER : My father came much earlier . And my family -- my mother joined with my oldest sister in \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    Paris\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        POPULATED_PLACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "-- to \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    Paris\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        POPULATED_PLACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       ". \n",
+       "\n",
+       " QUESTION : What brought your father to \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    Paris\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        POPULATED_PLACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "? \n",
+       "\n",
+       " ANSWER : I think the Army , to a certain extent , and a job , a good job . \n",
+       "\n",
+       " QUESTION : How large was your family ? \n",
+       "\n",
+       " ANSWER : Well , I had four sisters , two brothers , and my mother and father . I had a lot of family in \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    Poland\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        COUNTRY\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "; my grandparents , my aunts , my uncles , my cousins . \n",
+       "\n",
+       " QUESTION : But you were born in \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    Paris\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        POPULATED_PLACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       ". Did you go back to \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    Poland\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        COUNTRY\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "to visit ? \n",
+       "\n",
+       " ANSWER : Yes , my grandfather , we went to visit . \n",
+       "\n",
+       " QUESTION : And you stayed with large extended family there ? \n",
+       "\n",
+       " ANSWER : Yes , my grandfather did n't make my mother go \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    home\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        BUILDING\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       ", go back . \n",
+       "\n",
+       " QUESTION : He wanted her to stay ? \n",
+       "\n",
+       " ANSWER : Yes . \n",
+       "\n",
+       " QUESTION : Yeah . And so where were you in the order of children ? Were you the second ? \n",
+       "\n",
+       " ANSWER : Second . \n",
+       "\n",
+       " QUESTION : You were the second child ? \n",
+       "\n",
+       " ANSWER : Right . \n",
+       "\n",
+       " QUESTION : Yeah . And what kind of \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    neighborhood\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        POPULATED_PLACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "did you live in in \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    Paris\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        POPULATED_PLACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "? \n",
+       "\n",
+       " ANSWER : We lived in a mixed neighborhood . \n",
+       "\n",
+       " QUESTION : Non - Jews and -- ANSWER : Non - Jews and Jews . \n",
+       "\n",
+       " QUESTION : Uh - huh . And did you have non - Jewish friends ? \n",
+       "\n",
+       " ANSWER : Yes . Actually , I was -- when I came back from the \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    camp\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        POPULATED_PLACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       ", I lived with my friend I went to \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    school\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        BUILDING\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "with when we was six and a half years old . \n",
+       "\n",
+       " QUESTION : Uh - huh . \n",
+       "\n",
+       " ANSWER : And she was Catholic . \n",
+       "\n",
+       " QUESTION : Uh - huh . \n",
+       "\n",
+       " ANSWER : My Jewish family did n't have \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    room\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        INT_SPACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "for me . \n",
+       "\n",
+       " QUESTION : Uh - huh . Let 's talk a little bit about your schooling . When did you begin \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    school\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        BUILDING\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "? How old were you ? \n",
+       "\n",
+       " ANSWER : We begin \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    school\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        BUILDING\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "early in \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    France\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        COUNTRY\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       ". We go to first the maternelle , which was maybe about three years old . And the -- QUESTION : Is this a \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    public\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        BUILDING\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "\n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    school\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "</span>\n",
+       "? \n",
+       "\n",
+       " ANSWER : Yes . \n",
+       "\n",
+       " QUESTION : So there were Jews and non - Jews in your class ? \n",
+       "\n",
+       " ANSWER : Yes , Jews and non - Jews . But mostly there were non - Jews in my \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    school\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        BUILDING\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       ". And then first grade , we start -- I think I was six , six and a half . And when I was arrested , I was in lycee . And did n't finish my education , therefore . Q : Yeah . Yeah . \n",
+       "\n",
+       " ANSWER : Six years . \n",
+       "\n",
+       " QUESTION : Yeah . We 'll get to that in a few minutes . So you -- you went to a regular \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    public\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        BUILDING\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "\n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    school\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "</span>\n",
+       ". \n",
+       "\n",
+       " ANSWER : Uh - huh . \n",
+       "\n",
+       " QUESTION : That 's now what we call \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    elementary\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        BUILDING\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "\n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    school\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "</span>\n",
+       ". \n",
+       "\n",
+       " ANSWER : Yes . \n",
+       "\n",
+       " QUESTION : Was there any problems there ? You said 1 t was mixed . \n",
+       "\n",
+       " ANSWER : Yeah . No problem . \n",
+       "\n",
+       " QUESTION : No problems ? \n",
+       "\n",
+       " ANSWER : No . \n",
+       "\n",
+       " QUESTION : You were accepted ? \n",
+       "\n",
+       " ANSWER : Yes . No problem . On the contrary , our teachers liked us very much . We were five girls in the same \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    school\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        BUILDING\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       ", because we had \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    girls\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        BUILDING\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "\n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    '\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "</span>\n",
+       "\n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    schools\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "</span>\n",
+       "and \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    boys\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        BUILDING\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "\n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    '\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "</span>\n",
+       "\n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    schools\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "</span>\n",
+       ". It was not mixed \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    schools\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        BUILDING\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       ". And the -- my principal , when we were arrested , came every single day to see if there were news from us . So it was no problem at all . \n",
+       "\n",
+       " QUESTION : Uh - huh . What kind of \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    neighborhood\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        POPULATED_PLACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "did you live in ? Was it upper class or middle class neighborhood ? \n",
+       "\n",
+       " ANSWER : Probably middle class , I would say . \n",
+       "\n",
+       " QUESTION : Uh - huh . Did you live in a \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    house\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        BUILDING\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "or an \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    apartment\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        INT_SPACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "? \n",
+       "\n",
+       " ANSWER : A \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    house\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        BUILDING\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       ". \n",
+       "\n",
+       " QUESTION : And was it in the center of the \n",
+       "<span style=\"font-weight: bold; display: inline-block; position: relative; height: 60px;\">\n",
+       "    city\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "</span>\n",
+       "\n",
+       "    \n",
+       "<span style=\"background: #ddd; top: 40px; height: 4px; border-top-left-radius: 3px; border-bottom-left-radius: 3px; left: -1px; width: calc(100% + 2px); position: absolute;\">\n",
+       "    <span style=\"background: #ddd; z-index: 10; color: #000; top: -0.5em; padding: 2px 3px; position: absolute; font-size: 0.6em; font-weight: bold; line-height: 1; border-radius: 3px\">\n",
+       "        POPULATED_PLACE\n",
+       "    </span>\n",
+       "</span>\n",
+       "\n",
+       "\n",
+       "</span>\n",
+       "or the outskirts ? \n",
+       " </div></span>"
+      ],
+      "text/plain": [
+       "<IPython.core.display.HTML object>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "displacy.render(doc, style=\"span\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "holocaust",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.13"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

project.lock

@@ -0,0 +1,174 @@
+split:
+  cmd: python -m weasel run split
+  script:
+    - python scripts/split.py assets/annotated_data_spans.jsonl
+  deps:
+    - path: scripts/split.py
+      md5: 9983293e4982a7e989ba51456bfc992c
+  outs:
+    - path: assets/train.jsonl
+      md5: 584673b1949bfaaf9166c2212401f728
+    - path: assets/dev.jsonl
+      md5: 8dab61f2033edcfb8bbeb20a1454283d
+    - path: assets/test.jsonl
+      md5: 75ef0134ef3aa68ce76230fc6cc80dd4
+convert:
+  cmd: python -m weasel run convert
+  script:
+    - python scripts/convert.py en assets/train.jsonl corpus
+    - python scripts/convert.py en assets/dev.jsonl corpus
+    - python scripts/convert.py en assets/test.jsonl corpus
+  deps:
+    - path: assets/train.jsonl
+      md5: 584673b1949bfaaf9166c2212401f728
+    - path: assets/dev.jsonl
+      md5: 8dab61f2033edcfb8bbeb20a1454283d
+    - path: assets/test.jsonl
+      md5: 75ef0134ef3aa68ce76230fc6cc80dd4
+    - path: scripts/convert.py
+      md5: 8e1f918fad38cf8867e50977d21c0408
+  outs:
+    - path: corpus/train.spacy
+      md5: 90c3e968d70b05835fc8042d534b15e0
+    - path: corpus/dev.spacy
+      md5: 174b964498a65cbe84993e303c7c4793
+    - path: corpus/test.spacy
+      md5: b9e5adef7f8b78fe804037a8e59a4cf0
+train:
+  cmd: python -m weasel run train
+  script:
+    - python -m spacy train configs/config_trf.cfg --output training/  --paths.train
+      corpus/train.spacy --paths.dev corpus/dev.spacy  --training.eval_frequency 10  --training.patience
+      100  --gpu-id -1 --system.seed 0
+  deps:
+    - path: configs/config.cfg
+      md5: 479f0a665e528cc7f32ca449a55cb3d1
+    - path: corpus/train.spacy
+      md5: 90c3e968d70b05835fc8042d534b15e0
+    - path: corpus/dev.spacy
+      md5: 174b964498a65cbe84993e303c7c4793
+  outs:
+    - path: training/model-best
+      md5: 8d1843c1715af55bcd109ee6e7200afd
+evaluate:
+  cmd: python -m weasel run evaluate
+  script:
+    - python -m spacy evaluate training/model-best corpus/test.spacy --output training/metrics.json
+  deps:
+    - path: corpus/test.spacy
+      md5: b9e5adef7f8b78fe804037a8e59a4cf0
+    - path: training/model-best
+      md5: 8d1843c1715af55bcd109ee6e7200afd
+  outs:
+    - path: training/metrics.json
+      md5: 24152be0160e2a20e88a7577473b208b
+train-sm:
+  cmd: python -m weasel run train-sm
+  script:
+    - python -m spacy train configs/config_sm.cfg --output training/sm/  --paths.train
+      corpus/train.spacy --paths.dev corpus/dev.spacy  --training.patience 100  --gpu-id
+      -1 --system.seed 0
+  deps:
+    - path: configs/config.cfg
+      md5: 479f0a665e528cc7f32ca449a55cb3d1
+    - path: corpus/train.spacy
+      md5: 90c3e968d70b05835fc8042d534b15e0
+    - path: corpus/dev.spacy
+      md5: 174b964498a65cbe84993e303c7c4793
+  outs:
+    - path: training/model-best
+      md5: null
+evaluate-sm:
+  cmd: python -m weasel run evaluate-sm
+  script:
+    - python -m spacy evaluate training/sm/model-best corpus/test.spacy --output training/sm/metrics.json
+  deps:
+    - path: corpus/test.spacy
+      md5: b9e5adef7f8b78fe804037a8e59a4cf0
+    - path: training/sm/model-best
+      md5: 11cd6594ac09a0def1f01c51343576b0
+  outs:
+    - path: training/sm/metrics.json
+      md5: 6a8ffabb350c4f0b6262c6e9c45005af
+download-lg:
+  cmd: python -m weasel run download-lg
+  script:
+    - python -m spacy download en_core_web_lg
+  deps: []
+  outs: []
+train-lg:
+  cmd: python -m weasel run train-lg
+  script:
+    - python -m spacy train configs/config_lg.cfg --output training/lg/  --paths.train
+      corpus/train.spacy --paths.dev corpus/dev.spacy  --training.eval_frequency 50  --training.patience
+      0  --gpu-id -1  --initialize.vectors en_core_web_lg  --system.seed 0 --components.tok2vec.model.embed.include_static_vectors
+      true
+  deps:
+    - path: configs/config_lg.cfg
+      md5: 283a7f5e530c92c812e7666d52e539a6
+    - path: corpus/train.spacy
+      md5: 90c3e968d70b05835fc8042d534b15e0
+    - path: corpus/dev.spacy
+      md5: 174b964498a65cbe84993e303c7c4793
+  outs:
+    - path: training/model-best
+      md5: null
+evaluate-lg:
+  cmd: python -m weasel run evaluate-lg
+  script:
+    - python -m spacy evaluate training/lg/model-best corpus/test.spacy --output training/lg/metrics.json
+  deps:
+    - path: corpus/test.spacy
+      md5: b9e5adef7f8b78fe804037a8e59a4cf0
+    - path: training/lg/model-best
+      md5: 8e296e41d1ab48052ec122eed9154887
+  outs:
+    - path: training/lg/metrics.json
+      md5: 4efed860818658f1533fe204da7c169e
+download-md:
+  cmd: python -m weasel run download-md
+  script:
+    - python -m spacy download en_core_web_md
+  deps: []
+  outs: []
+evaluate-md:
+  cmd: python -m weasel run evaluate-md
+  script:
+    - python -m spacy evaluate training/md/model-best corpus/test.spacy --output training/md/metrics.json
+  deps:
+    - path: corpus/test.spacy
+      md5: b9e5adef7f8b78fe804037a8e59a4cf0
+    - path: training/md/model-best
+      md5: 2725d583e50534609145708af72679bc
+  outs:
+    - path: training/md/metrics.json
+      md5: b92630a9072590b0b9aef4393ca19106
+train-md:
+  cmd: python -m weasel run train-md
+  script:
+    - python -m spacy train configs/config_md.cfg --output training/md/  --paths.train
+      corpus/train.spacy --paths.dev corpus/dev.spacy  --training.eval_frequency 50  --training.patience
+      0  --gpu-id -1  --initialize.vectors en_core_web_md  --system.seed 0 --components.tok2vec.model.embed.include_static_vectors
+      true
+  deps:
+    - path: configs/config_md.cfg
+      md5: b641ff9a6ba2162be49f940eeade2004
+    - path: corpus/train.spacy
+      md5: 90c3e968d70b05835fc8042d534b15e0
+    - path: corpus/dev.spacy
+      md5: 174b964498a65cbe84993e303c7c4793
+  outs:
+    - path: training/model-best
+      md5: null
+build-table:
+  cmd: python -m weasel run build-table
+  script:
+    - python scripts/build-table.py
+  deps: []
+  outs: []
+readme:
+  cmd: python -m weasel run readme
+  script:
+    - python scripts/readme.py
+  deps: []
+  outs: []

project.yml

@@ -0,0 +1,268 @@
+title: "Demo spancat in a new pipeline (Span Categorization)"
+description: "A minimal demo spancat project for spaCy v3"
+
+# Variables can be referenced across the project.yml using ${vars.var_name}
+vars:
+  name: "placing_holocaust"
+  lang: "en"
+  annotations_file: "annotated_data_spans.jsonl"
+  train: "train"
+  dev: "dev"
+  test: "test"
+  version: "0.0.1"
+  # Set a random seed
+  seed: 0
+  # Set your GPU ID, -1 is CPU
+  gpu_id: -1
+  vectors_model_md: "en_core_web_md"
+  vectors_model_lg: "en_core_web_lg"
+
+# These are the directories that the project needs. The project CLI will make
+# sure that they always exist.
+directories: ["assets", "corpus", "configs", "training", "scripts", "packages"]
+
+# Assets that should be downloaded or available in the directory. We're shipping
+# them with the project, so they won't have to be downloaded.
+assets:
+  - dest: "assets/train.json"
+    description: "Demo training data adapted from the `ner_demo` project"
+  - dest: "assets/dev.json"
+    description: "Demo development data"
+
+# Workflows are sequences of commands (see below) executed in order. You can
+# run them via "spacy project run [workflow]". If a commands's inputs/outputs
+# haven't changed, it won't be re-run.
+workflows:
+  all-sm-sents:
+    - convert-sents
+    - split
+    - create-config-sm
+    - train-sm
+    - evaluate-sm
+  # all-trf:
+  #   - download
+  #   - convert
+  #   - create-config
+  #   - train-with-vectors
+  #   - evaluate
+
+# Project commands, specified in a style similar to CI config files (e.g. Azure
+# pipelines). The name is the command name that lets you trigger the command
+# via "spacy project run [command] [path]". The help message is optional and
+# shown when executing "spacy project run [optional command] [path] --help".
+commands:
+
+#### DOWNLOADING VECTORS #####
+  - name: "download-lg"
+    help: "Download a spaCy model with pretrained vectors"
+    script:
+      - "python -m spacy download ${vars.vectors_model_lg}"
+
+  - name: "download-md"
+    help: "Download a spaCy model with pretrained vectors"
+    script:
+      - "python -m spacy download ${vars.vectors_model_md}"
+
+#### PREPROCESSING #####
+  - name: "convert"
+    help: "Convert the data to spaCy's binary format"
+    script:
+      - "python scripts/convert.py ${vars.lang} assets/${vars.train}.jsonl corpus"
+      - "python scripts/convert.py ${vars.lang} assets/${vars.dev}.jsonl corpus"
+      - "python scripts/convert.py ${vars.lang} assets/${vars.test}.jsonl corpus"
+    deps:
+      - "assets/${vars.train}.jsonl"
+      - "assets/${vars.dev}.jsonl"
+      - "assets/${vars.test}.jsonl"
+      - "scripts/convert.py"
+    outputs:
+      - "corpus/train.spacy"
+      - "corpus/dev.spacy"
+      - "corpus/test.spacy"
+
+  - name: "convert-sents"
+    help: "Convert the data to to sentences before converting to spaCy's binary format"
+    script:
+      - "python scripts/convert_sents.py ${vars.lang} assets/${vars.train}.jsonl corpus"
+      - "python scripts/convert_sents.py ${vars.lang} assets/${vars.dev}.jsonl corpus"
+      - "python scripts/convert_sents.py ${vars.lang} assets/${vars.test}.jsonl corpus"
+    deps:
+      - "assets/${vars.train}.jsonl"
+      - "assets/${vars.dev}.jsonl"
+      - "assets/${vars.test}.jsonl"
+      - "scripts/convert.py"
+    outputs:
+      - "corpus/train.spacy"
+      - "corpus/dev.spacy"
+      - "corpus/test.spacy"
+
+  - name: "split"
+    help: "Split data into train/dev/test sets"
+    script:
+      - "python scripts/split.py assets/${vars.annotations_file}"
+    deps:
+      - "scripts/split.py"
+    outputs:
+      - "assets/train.jsonl"
+      - "assets/dev.jsonl"
+      - "assets/test.jsonl"
+
+
+
+#### CONFIG CREATIONS #####
+
+  - name: "create-config-sm"
+    help: "Create a new config with a spancat pipeline component"
+    script:
+      - "python -m spacy init fill-config configs/base_config_sm.cfg configs/config_sm.cfg"
+    deps:
+      - configs/base_config_sm.cfg
+    outputs:
+      - "configs/config.cfg"
+
+
+#### TRAINING #####
+
+### small ###
+  - name: "train-sm"
+    help: "Train the spancat model"
+    script:
+      - >-
+        python -m spacy train configs/config_sm.cfg --output training/sm/ 
+        --paths.train corpus/train.spacy --paths.dev corpus/dev.spacy
+         --training.eval_frequency 50 
+        --training.patience 0 
+        --gpu-id ${vars.gpu_id}
+        --system.seed ${vars.seed}
+    deps:
+      - "configs/config_lg.cfg"
+      - "corpus/train.spacy"
+      - "corpus/dev.spacy"
+    outputs:
+      - "training/model-best"
+
+
+### medium ###
+  - name: "train-md"
+    help: "Train the spancat model with vectors"
+    script:
+      - >-
+        python -m spacy train configs/config_md.cfg --output training/md/ 
+        --paths.train corpus/train.spacy --paths.dev corpus/dev.spacy 
+        --training.eval_frequency 50 
+        --training.patience 0 
+        --gpu-id ${vars.gpu_id} 
+        --initialize.vectors ${vars.vectors_model_md} 
+        --system.seed ${vars.seed}
+        --components.tok2vec.model.embed.include_static_vectors true
+    deps:
+      - "configs/config_md.cfg"
+      - "corpus/train.spacy"
+      - "corpus/dev.spacy"
+    outputs:
+      - "training/model-best"
+
+
+### large ###
+  - name: "train-lg"
+    help: "Train the spancat model with vectors"
+    script:
+      - >-
+        python -m spacy train configs/config_lg.cfg --output training/lg/ 
+        --paths.train corpus/train.spacy --paths.dev corpus/dev.spacy 
+        --training.eval_frequency 50 
+        --training.patience 0 
+        --gpu-id ${vars.gpu_id} 
+        --initialize.vectors ${vars.vectors_model_lg} 
+        --system.seed ${vars.seed}
+        --components.tok2vec.model.embed.include_static_vectors true
+    deps:
+      - "configs/config_lg.cfg"
+      - "corpus/train.spacy"
+      - "corpus/dev.spacy"
+    outputs:
+      - "training/model-best"
+
+
+### transformer ###
+  - name: "train-trf"
+    help: "Train the spancat model"
+    script:
+      - >-
+        python -m spacy train configs/config_trf.cfg --output training/trf/ 
+        --paths.train corpus/train.spacy --paths.dev corpus/dev.spacy 
+        --training.patience 100
+        --gpu-id ${vars.gpu_id}
+        --system.seed ${vars.seed}
+    deps:
+      - "configs/config.cfg"
+      - "corpus/train.spacy"
+      - "corpus/dev.spacy"
+    outputs:
+      - "training/model-best"
+
+
+#### EVALUATION #####
+
+### small ###
+  - name: "evaluate-sm"
+    help: "Evaluate the model and export metrics"
+    script:
+      - "python -m spacy evaluate training/sm/model-best corpus/test.spacy --output training/sm/metrics.json"
+    deps:
+      - "corpus/test.spacy"
+      - "training/sm/model-best"
+    outputs:
+      - "training/sm/metrics.json"
+
+### medium ###
+
+  - name: "evaluate-md"
+    help: "Evaluate the model and export metrics"
+    script:
+      - "python -m spacy evaluate training/md/model-best corpus/test.spacy --output training/md/metrics.json"
+    deps:
+      - "corpus/test.spacy"
+      - "training/md/model-best"
+    outputs:
+      - "training/md/metrics.json"
+
+### large ###
+  - name: "evaluate-lg"
+    help: "Evaluate the model and export metrics"
+    script:
+      - "python -m spacy evaluate training/lg/model-best corpus/test.spacy --output training/lg/metrics.json"
+    deps:
+      - "corpus/test.spacy"
+      - "training/lg/model-best"
+    outputs:
+      - "training/lg/metrics.json"
+
+
+#### PACKAGING #####
+
+  - name: "build-table"
+    help: "builds a nice table from the metrics for README.md"
+    script:
+      - "python scripts/build-table.py"
+
+  - name: "readme"
+    help: "builds a nice table from the metrics for README.md"
+    script:
+      - "python scripts/readme.py"
+
+  - name: package
+    help: "Package the trained model as a pip package"
+    script:
+      - "python -m spacy package training/model-best packages --name ${vars.name} --version ${vars.version} --force"
+    deps:
+      - "training/model-best"
+    outputs_no_cache:
+      - "packages/${vars.lang}_${vars.name}-${vars.version}/dist/${vars.lang}_${vars.name}-${vars.version}.tar.gz"
+
+  - name: clean
+    help: "Remove intermediary directories"
+    script:
+      - "rm -rf corpus/*"
+      - "rm -rf training/*"
+      - "rm -rf metrics/*"

requirements.txt

@@ -0,0 +1,3 @@
+scikit-learn
+tabulate
+pandas

scripts/build-table.py

@@ -0,0 +1,68 @@
+import os
+import json
+import pandas as pd
+from tabulate import tabulate
+import typer
+
+def generate_detailed_markdown_chart():
+    # Directories for each model type in the desired order
+    model_dirs = [
+        ('small', 'training/sm'),
+        ('medium', 'training/md'),
+        ('large', 'training/lg'),
+        ('transformer', 'training/trf')
+    ]
+
+    # DataFrame to hold the overall data
+    overall_df = pd.DataFrame(columns=['Model', 'Precision', 'Recall', 'F-Score'])
+    
+    # DataFrame to hold the per-type data
+    per_type_df = pd.DataFrame(columns=['Model', 'Label', 'Precision', 'Recall', 'F-Score'])
+
+    for model_name, dir_path in model_dirs:
+        metrics_file = os.path.join(dir_path, 'metrics.json')
+
+        # Check if the file exists
+        if os.path.exists(metrics_file):
+            with open(metrics_file, 'r') as file:
+                metrics = json.load(file)
+                # Extract overall metrics
+                overall_df = overall_df.append({
+                    'Model': model_name.capitalize(),
+                    'Precision': round(metrics['spans_sc_p'] * 100, 1),
+                    'Recall': round(metrics['spans_sc_r'] * 100, 1),
+                    'F-Score': round(metrics['spans_sc_f'] * 100, 1)
+                }, ignore_index=True)
+
+                # Extract per-type metrics
+                for label, scores in metrics.get('spans_sc_per_type', {}).items():
+                    per_type_df = per_type_df.append({
+                        'Model': model_name.capitalize(),
+                        'Label': label,
+                        'Precision': round(scores['p'] * 100, 1),
+                        'Recall': round(scores['r'] * 100, 1),
+                        'F-Score': round(scores['f'] * 100, 1)
+                    }, ignore_index=True)
+
+    # Define the order for models
+    model_order = ['Small', 'Medium', 'Large', 'Transformer']
+    per_type_df['Model'] = pd.Categorical(per_type_df['Model'], categories=model_order, ordered=True)
+
+    # Sort the per_type_df first by Label, then by Model
+    per_type_df.sort_values(by=['Label', 'Model'], inplace=True)
+
+    # Convert the DataFrames to Markdown
+    overall_markdown = tabulate(overall_df, headers='keys', tablefmt='pipe', showindex=False)
+    per_type_markdown = tabulate(per_type_df, headers='keys', tablefmt='pipe', showindex=False)
+
+    # Write the Markdown tables to a file
+    with open('model_comparison.md', 'w') as md_file:
+        md_file.write("# Overall Model Performance\n")
+        md_file.write(overall_markdown)
+        md_file.write("\n\n# Performance per Label\n")
+        md_file.write(per_type_markdown)
+
+    print("Markdown chart created as 'model_comparison.md'")
+
+if __name__ == "__main__":
+    typer.run(generate_detailed_markdown_chart)

scripts/convert.py

@@ -0,0 +1,41 @@
+"""Convert entity annotation from spaCy v2 TRAIN_DATA format to spaCy v3 .spacy format."""
+import srsly
+import typer
+import warnings
+from pathlib import Path
+import spacy
+from spacy.tokens import DocBin
+
+def convert(lang: str, input_paths: list[Path], output_dir: Path, spans_key: str = "sc"):
+    nlp = spacy.blank(lang)
+    nlp.add_pipe("sentencizer")
+    
+    # Ensure output directory exists
+    output_dir.mkdir(parents=True, exist_ok=True)
+
+    # Process each input file
+    for input_path in input_paths:
+        print(input_path)
+        doc_bin = DocBin()
+        for annotation in srsly.read_jsonl(input_path):
+            text = annotation["text"]
+            doc = nlp.make_doc(text)
+            spans = []
+            for item in annotation["spans"]:
+                start = item["start"]
+                end = item["end"]
+                label = item["label"]
+                span = doc.char_span(start, end, label=label)
+                if span is None:
+                    msg = f"Skipping entity [{start}, {end}, {label}] in the following text because the character span '{doc.text[start:end]}' does not align with token boundaries."
+                    warnings.warn(msg)
+                else:
+                    spans.append(span)
+            doc.spans[spans_key] = spans
+            doc_bin.add(doc)
+        # Write to output file in the specified directory
+        output_file = output_dir / f"{input_path.stem}.spacy"
+        doc_bin.to_disk(output_file)
+
+if __name__ == "__main__":
+    typer.run(convert)

scripts/convert_sents.py

@@ -0,0 +1,58 @@
+import srsly
+import typer
+import warnings
+from pathlib import Path
+import spacy
+from spacy.tokens import DocBin
+
+def convert(lang: str, input_paths: list[Path], output_dir: Path, spans_key: str = "sc"):
+    nlp = spacy.blank(lang)
+    nlp.add_pipe("sentencizer")
+    
+    # Ensure output directory exists
+    output_dir.mkdir(parents=True, exist_ok=True)
+
+    total_sentences = 0
+
+    # Process each input file
+    for input_path in input_paths:
+        print(f"Processing file: {input_path}")
+        doc_bin = DocBin()
+        
+        for annotation in srsly.read_jsonl(input_path):
+            text = annotation["text"]
+            doc = nlp(text)  # Process the document to split into sentences
+            
+            for sent in doc.sents:
+                # Create a new Doc for the sentence
+                sent_doc = nlp.make_doc(sent.text)
+                spans = []
+                for item in annotation["spans"]:
+                    # Adjust span start and end for the sentence
+                    start = item["start"] - sent.start_char
+                    end = item["end"] - sent.start_char
+                    label = item["label"]
+
+                    # Only consider spans that are within the sentence
+                    if start >= 0 and end <= len(sent.text):
+                        span = sent_doc.char_span(start, end, label=label, alignment_mode="contract")
+                        if span is None:
+                            msg = f"Skipping entity [{start}, {end}, {label}] in the following text because the character span '{sent.text[start:end]}' does not align with token boundaries."
+                            warnings.warn(msg)
+                        else:
+                            spans.append(span)
+                
+                # Add sentence to DocBin only if it contains spans
+                if spans:
+                    sent_doc.spans[spans_key] = spans
+                    doc_bin.add(sent_doc)
+                    total_sentences += 1
+
+        # Write to output file in the specified directory
+        output_file = output_dir / f"{input_path.stem}.spacy"
+        doc_bin.to_disk(output_file)
+    
+    print(f"Total sentences with spans: {total_sentences}")
+
+if __name__ == "__main__":
+    typer.run(convert)

scripts/readme.py

@@ -0,0 +1,82 @@
+import os
+import json
+import pandas as pd
+from tabulate import tabulate
+import typer
+
+def create_readme_for_model(model_dir: str, project_url: str):
+    # Path to the metrics and meta files
+    metrics_file = os.path.join(model_dir, 'metrics.json')
+    meta_file = os.path.join(model_dir, 'model-best', 'meta.json')
+
+    # DataFrame for the model's overall performance metrics
+    overall_df = pd.DataFrame(columns=['Metric', 'Value'])
+
+    # DataFrame for the model's per-label performance metrics
+    per_label_df = pd.DataFrame(columns=['Label', 'Precision', 'Recall', 'F-Score'])
+
+    # Read and add metrics data
+    if os.path.exists(metrics_file):
+        with open(metrics_file, 'r') as file:
+            metrics = json.load(file)
+            overall_df = overall_df.append({'Metric': 'Precision', 'Value': round(metrics['spans_sc_p'] * 100, 1)}, ignore_index=True)
+            overall_df = overall_df.append({'Metric': 'Recall', 'Value': round(metrics['spans_sc_r'] * 100, 1)}, ignore_index=True)
+            overall_df = overall_df.append({'Metric': 'F-Score', 'Value': round(metrics['spans_sc_f'] * 100, 1)}, ignore_index=True)
+
+            # Extract and add per-type metrics
+            for label, scores in metrics.get('spans_sc_per_type', {}).items():
+                per_label_df = per_label_df.append({
+                    'Label': label,
+                    'Precision': round(scores['p'] * 100, 1),
+                    'Recall': round(scores['r'] * 100, 1),
+                    'F-Score': round(scores['f'] * 100, 1)
+                }, ignore_index=True)
+
+    # Sort the per_label_df by Label
+    per_label_df.sort_values(by='Label', inplace=True)
+
+    # Convert the DataFrames to Markdown tables
+    overall_markdown = tabulate(overall_df, headers='keys', tablefmt='pipe', showindex=False)
+    per_label_markdown = tabulate(per_label_df, headers='keys', tablefmt='pipe', showindex=False)
+
+    # Read meta.json file
+    meta_info = ""
+    if os.path.exists(meta_file):
+        with open(meta_file, 'r') as file:
+            meta_data = json.load(file)
+            for key, value in meta_data.items():
+                meta_info += f"- **{key}**: {value}\n"
+
+    # README content
+    readme_content = f"""
+# Placing the Holocaust spaCy Model - {os.path.basename(model_dir).capitalize()}
+
+This is a spaCy model trained as part of the placingholocaust spaCy project. Training and evaluation code, along with the dataset, can be found at the following URL: [Placingholocaust SpaCy Project]({project_url})
+
+## Model Performance
+{overall_markdown}
+
+## Performance per Label
+{per_label_markdown}
+
+## Meta Information
+{meta_info}
+"""
+
+    # Write the README content to a file
+    readme_file = os.path.join(model_dir, 'README.md')
+    with open(readme_file, 'w') as file:
+        file.write(readme_content)
+
+    print(f"README created in {model_dir}")
+
+def create_all_readmes(project_url: str):
+    # Directories for each model type
+    model_dirs = ['training/sm', 'training/md', 'training/lg', 'training/trf']
+
+    for dir in model_dirs:
+        create_readme_for_model(dir, project_url)
+
+if __name__ == "__main__":
+    project_url = "https://huggingface.co/datasets/placingholocaust/spacy-project"
+    typer.run(lambda: create_all_readmes(project_url))

scripts/split.py

@@ -0,0 +1,54 @@
+import toml
+from collections import Counter
+import srsly
+import typer
+from sklearn.model_selection import train_test_split
+from pathlib import Path
+
+def count_labels(data):
+    label_counter = Counter()
+    for annotation in data:
+        labels = [span['label'] for span in annotation['spans']]
+        label_counter.update(labels)
+    return label_counter
+
+def split_data(input_file: Path, train_ratio: float = 0.7, dev_ratio: float = 0.15, 
+               train_output: Path = Path("assets/train.jsonl"), 
+               dev_output: Path = Path("assets/dev.jsonl"), 
+               test_output: Path = Path("assets/test.jsonl"), 
+               random_state: int = 1):
+    # Read data from JSONL
+    data = list(srsly.read_jsonl(input_file))
+
+    # Split data
+    test_ratio = 1 - train_ratio - dev_ratio
+    train_data, temp_data = train_test_split(data, test_size=(dev_ratio + test_ratio), random_state=random_state)
+    dev_data, test_data = train_test_split(temp_data, test_size=test_ratio/(dev_ratio + test_ratio), random_state=random_state)
+
+    # Count labels in each dataset
+    train_labels = count_labels(train_data)
+    dev_labels = count_labels(dev_data)
+    test_labels = count_labels(test_data)
+
+    # Write split data to JSONL files
+    srsly.write_jsonl(train_output, train_data)
+    srsly.write_jsonl(dev_output, dev_data)
+    srsly.write_jsonl(test_output, test_data)
+
+    # Combine label counts into a dictionary
+    all_labels = sorted(set(train_labels.keys()) | set(dev_labels.keys()) | set(test_labels.keys()))
+    annotations_data = {label: {'Train': train_labels.get(label, 0), 'Dev': dev_labels.get(label, 0), 'Test': test_labels.get(label, 0)} for label in all_labels}
+
+    # Print the table
+    print(f"{'Label':<20}{'Train':<10}{'Dev':<10}{'Test':<10}")
+    for label in all_labels:
+        print(f"{label:<20}{train_labels.get(label, 0):<10}{dev_labels.get(label, 0):<10}{test_labels.get(label, 0):<10}")
+
+    # Save the table in annotations.toml
+    with open("annotations.toml", "w") as toml_file:
+        toml.dump(annotations_data, toml_file)
+
+    print("Annotations data saved in annotations.toml")
+
+if __name__ == "__main__":
+    typer.run(split_data)

spacy-project.md

@@ -0,0 +1,53 @@
+<!-- WEASEL: AUTO-GENERATED DOCS START (do not remove) -->
+
+# 🪐 Weasel Project: Demo spancat in a new pipeline (Span Categorization)
+
+A minimal demo spancat project for spaCy v3
+
+## 📋 project.yml
+
+The [`project.yml`](project.yml) defines the data assets required by the
+project, as well as the available commands and workflows. For details, see the
+[Weasel documentation](https://github.com/explosion/weasel).
+
+### ⏯ Commands
+
+The following commands are defined by the project. They
+can be executed using [`weasel run [name]`](https://github.com/explosion/weasel/tree/main/docs/cli.md#rocket-run).
+Commands are only re-run if their inputs have changed.
+
+| Command | Description |
+| --- | --- |
+| `download` | Download a spaCy model with pretrained vectors |
+| `convert` | Convert the data to spaCy's binary format |
+| `create-config` | Create a new config with a spancat pipeline component |
+| `train` | Train the spancat model |
+| `train-with-vectors` | Train the spancat model with vectors |
+| `evaluate` | Evaluate the model and export metrics |
+| `package` | Package the trained model as a pip package |
+| `clean` | Remove intermediary directories |
+
+### ⏭ Workflows
+
+The following workflows are defined by the project. They
+can be executed using [`weasel run [name]`](https://github.com/explosion/weasel/tree/main/docs/cli.md#rocket-run)
+and will run the specified commands in order. Commands are only re-run if their
+inputs have changed.
+
+| Workflow | Steps |
+| --- | --- |
+| `all` | `convert` &rarr; `create-config` &rarr; `train` &rarr; `evaluate` |
+| `all-vectors` | `download` &rarr; `convert` &rarr; `create-config` &rarr; `train-with-vectors` &rarr; `evaluate` |
+
+### 🗂 Assets
+
+The following assets are defined by the project. They can
+be fetched by running [`weasel assets`](https://github.com/explosion/weasel/tree/main/docs/cli.md#open_file_folder-assets)
+in the project directory.
+
+| File | Source | Description |
+| --- | --- | --- |
+| [`assets/train.json`](assets/train.json) | Local | Demo training data adapted from the `ner_demo` project |
+| [`assets/dev.json`](assets/dev.json) | Local | Demo development data |
+
+<!-- WEASEL: AUTO-GENERATED DOCS END (do not remove) -->