Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
|
@@ -5,24 +5,35 @@ import pdfplumber
|
|
| 5 |
import pandas as pd
|
| 6 |
import time
|
| 7 |
from cnocr import CnOcr
|
| 8 |
-
|
| 9 |
|
| 10 |
-
word_embedding_model = models.Transformer('uer/sbert-base-chinese-nli', do_lower_case=True) # BERT模型
|
| 11 |
-
pooling_model = models.Pooling(word_embedding_model.get_word_embedding_dimension(), pooling_mode='cls') # 取cls向量作为句向量
|
| 12 |
-
embedder = SentenceTransformer(modules=[word_embedding_model, pooling_model]) # 定义模型
|
| 13 |
ocr = CnOcr() # 初始化ocr模型
|
| 14 |
-
chat_url = 'https://souljoy-my-api.hf.space/chatgpt' # 你的url
|
| 15 |
-
headers = {
|
| 16 |
-
'Content-Type': 'application/json',
|
| 17 |
-
} # 你的headers
|
| 18 |
history_max_len = 500 # 机器人记忆的最大长度
|
| 19 |
all_max_len = 3000 # 输入的最大长度
|
| 20 |
|
| 21 |
|
| 22 |
-
def
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 23 |
texts = doc.split('\n') # 按行切分
|
| 24 |
-
emb_list =
|
| 25 |
-
|
|
|
|
| 26 |
return texts, emb_list, gr.Textbox.update(visible=True), gr.Button.update(visible=True), gr.Markdown.update(
|
| 27 |
value="""操作说明 step 3:PDF解析提交成功! 🙋 可以开始对话啦~"""), gr.Chatbot.update(visible=True)
|
| 28 |
|
|
@@ -36,9 +47,17 @@ def get_response(open_ai_key, msg, bot, doc_text_list, doc_embeddings): # 获
|
|
| 36 |
now_len += len(bot[i][0]) + len(bot[i][1]) # 更新当前长度
|
| 37 |
his_bg = i # 更新历史记录的起始位置
|
| 38 |
history = [] if his_bg == -1 else bot[his_bg:] # 获取历史记录
|
| 39 |
-
query_embedding =
|
| 40 |
-
cos_scores =
|
| 41 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 42 |
score_index.sort(key=lambda x: x[0], reverse=True) # 按相似度排序
|
| 43 |
print('score_index:\n', score_index)
|
| 44 |
index_set, sub_doc_list = set(), [] # 用于存储最终的索引和文档
|
|
@@ -75,14 +94,24 @@ def get_response(open_ai_key, msg, bot, doc_text_list, doc_embeddings): # 获
|
|
| 75 |
messages.append({"role": "user", "content": his[0]}) # 加入用户的历史记录
|
| 76 |
messages.append({"role": "assistant", "content": his[1]}) # 加入机器人的历史记录
|
| 77 |
messages.append({"role": "user", "content": msg}) # 加入用户的当前输入
|
| 78 |
-
|
| 79 |
-
|
| 80 |
-
|
| 81 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 82 |
data=json.dumps(data),
|
| 83 |
headers=headers
|
| 84 |
-
)
|
| 85 |
-
res = result.json()['content']
|
| 86 |
bot.append([msg, res]) # 加入历史记录
|
| 87 |
return bot[max(0, len(bot) - 3):] # 返回最近3轮的历史记录
|
| 88 |
|
|
@@ -124,7 +153,7 @@ def up_file(files): # 上传文件
|
|
| 124 |
print(doc_text_list)
|
| 125 |
return gr.Textbox.update(value='\n'.join(doc_text_list), visible=True), gr.Button.update(
|
| 126 |
visible=True), gr.Markdown.update(
|
| 127 |
-
value="操作说明 step 2:确认PDF
|
| 128 |
|
| 129 |
|
| 130 |
with gr.Blocks() as demo:
|
|
@@ -134,7 +163,8 @@ with gr.Blocks() as demo:
|
|
| 134 |
file = gr.File(file_types=['.pdf'], label='点击上传PDF,进行解析(支持多文档、表格、OCR)',
|
| 135 |
file_count='multiple') # 支持多文档、表格、OCR
|
| 136 |
txt = gr.Textbox(label='PDF解析结果', visible=False) # PDF解析结果
|
| 137 |
-
|
|
|
|
| 138 |
doc_text_state = gr.State([]) # 存储PDF解析结果
|
| 139 |
doc_emb_state = gr.State([]) # 存储PDF解析结果的embedding
|
| 140 |
with gr.Column():
|
|
@@ -144,8 +174,9 @@ with gr.Blocks() as demo:
|
|
| 144 |
with gr.Row():
|
| 145 |
chat_bu = gr.Button(value='发送', visible=False) # 发送按钮
|
| 146 |
|
| 147 |
-
file.change(up_file, [file], [txt,
|
| 148 |
-
|
|
|
|
| 149 |
chat_bu.click(get_response, [open_ai_key, msg_txt, chat_bot, doc_text_state, doc_emb_state], [chat_bot]) # 发送消息
|
| 150 |
|
| 151 |
if __name__ == "__main__":
|
|
|
|
| 5 |
import pandas as pd
|
| 6 |
import time
|
| 7 |
from cnocr import CnOcr
|
| 8 |
+
import numpy as np
|
| 9 |
|
|
|
|
|
|
|
|
|
|
| 10 |
ocr = CnOcr() # 初始化ocr模型
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
history_max_len = 500 # 机器人记忆的最大长度
|
| 12 |
all_max_len = 3000 # 输入的最大长度
|
| 13 |
|
| 14 |
|
| 15 |
+
def get_text_emb(open_ai_key, text):
|
| 16 |
+
url = 'https://api.openai.com/v1/embeddings'
|
| 17 |
+
headers = {
|
| 18 |
+
'Content-Type': 'application/json',
|
| 19 |
+
'Authorization': 'Bearer ' + open_ai_key
|
| 20 |
+
}
|
| 21 |
+
data = {
|
| 22 |
+
"model": "text-embedding-ada-002",
|
| 23 |
+
"input": text
|
| 24 |
+
}
|
| 25 |
+
result = requests.post(url=url,
|
| 26 |
+
data=json.dumps(data),
|
| 27 |
+
headers=headers
|
| 28 |
+
)
|
| 29 |
+
return result.json()['data'][0]['embedding']
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
def doc_index_self(open_ai_key, doc): # 文档向量化
|
| 33 |
texts = doc.split('\n') # 按行切分
|
| 34 |
+
emb_list = []
|
| 35 |
+
for text in texts:
|
| 36 |
+
emb_list.append(get_text_emb(open_ai_key, text))
|
| 37 |
return texts, emb_list, gr.Textbox.update(visible=True), gr.Button.update(visible=True), gr.Markdown.update(
|
| 38 |
value="""操作说明 step 3:PDF解析提交成功! 🙋 可以开始对话啦~"""), gr.Chatbot.update(visible=True)
|
| 39 |
|
|
|
|
| 47 |
now_len += len(bot[i][0]) + len(bot[i][1]) # 更新当前长度
|
| 48 |
his_bg = i # 更新历史记录的起始位置
|
| 49 |
history = [] if his_bg == -1 else bot[his_bg:] # 获取历史记录
|
| 50 |
+
query_embedding = get_text_emb(open_ai_key, msg) # 获取输入的向量
|
| 51 |
+
cos_scores = [] # 用于存储相似度
|
| 52 |
+
|
| 53 |
+
def cos_sim(a, b):
|
| 54 |
+
return np.dot(a, b) / (np.linalg.norm(a) * np.linalg.norm(b))
|
| 55 |
+
|
| 56 |
+
for doc_embedding in doc_embeddings: # 遍历文档向量
|
| 57 |
+
cos_scores.append(cos_sim(query_embedding, doc_embedding)) # 计算相似度
|
| 58 |
+
score_index = [] # 用于存储相似度和索引对应
|
| 59 |
+
for i in range(len(cos_scores)): # 遍历相似度
|
| 60 |
+
score_index.append((cos_scores[i], i)) # 加入相似度和索引对应
|
| 61 |
score_index.sort(key=lambda x: x[0], reverse=True) # 按相似度排序
|
| 62 |
print('score_index:\n', score_index)
|
| 63 |
index_set, sub_doc_list = set(), [] # 用于存储最终的索引和文档
|
|
|
|
| 94 |
messages.append({"role": "user", "content": his[0]}) # 加入用户的历史记录
|
| 95 |
messages.append({"role": "assistant", "content": his[1]}) # 加入机器人的历史记录
|
| 96 |
messages.append({"role": "user", "content": msg}) # 加入用户的当前输入
|
| 97 |
+
|
| 98 |
+
url = 'https://api.openai.com/v1/chat/completions'
|
| 99 |
+
|
| 100 |
+
data = {
|
| 101 |
+
"model": "gpt-3.5-turbo",
|
| 102 |
+
"messages": messages
|
| 103 |
+
}
|
| 104 |
+
print("data = \n", data)
|
| 105 |
+
|
| 106 |
+
headers = {
|
| 107 |
+
'Content-Type': 'application/json',
|
| 108 |
+
'Authorization': 'Bearer ' + open_ai_key
|
| 109 |
+
}
|
| 110 |
+
result = requests.post(url=url,
|
| 111 |
data=json.dumps(data),
|
| 112 |
headers=headers
|
| 113 |
+
)
|
| 114 |
+
res = str(result.json()['choices'][0]['message']['content']).strip()
|
| 115 |
bot.append([msg, res]) # 加入历史记录
|
| 116 |
return bot[max(0, len(bot) - 3):] # 返回最近3轮的历史记录
|
| 117 |
|
|
|
|
| 153 |
print(doc_text_list)
|
| 154 |
return gr.Textbox.update(value='\n'.join(doc_text_list), visible=True), gr.Button.update(
|
| 155 |
visible=True), gr.Markdown.update(
|
| 156 |
+
value="操作说明 step 2:确认PDF解析结果(可修正),点击“建立索引”,随后进行对话")
|
| 157 |
|
| 158 |
|
| 159 |
with gr.Blocks() as demo:
|
|
|
|
| 163 |
file = gr.File(file_types=['.pdf'], label='点击上传PDF,进行解析(支持多文档、表格、OCR)',
|
| 164 |
file_count='multiple') # 支持多文档、表格、OCR
|
| 165 |
txt = gr.Textbox(label='PDF解析结果', visible=False) # PDF解析结果
|
| 166 |
+
index_self_bu = gr.Button(value='建立索引(by self)', visible=False) #
|
| 167 |
+
index_llama_bu = gr.Button(value='建立索引(by llama_index)', visible=False) #
|
| 168 |
doc_text_state = gr.State([]) # 存储PDF解析结果
|
| 169 |
doc_emb_state = gr.State([]) # 存储PDF解析结果的embedding
|
| 170 |
with gr.Column():
|
|
|
|
| 174 |
with gr.Row():
|
| 175 |
chat_bu = gr.Button(value='发送', visible=False) # 发送按钮
|
| 176 |
|
| 177 |
+
file.change(up_file, [file], [txt, index_self_bu, md]) # 上传文件
|
| 178 |
+
index_self_bu.click(doc_index_self, [txt],
|
| 179 |
+
[doc_text_state, doc_emb_state, msg_txt, chat_bu, md, chat_bot]) # 提交解析结果
|
| 180 |
chat_bu.click(get_response, [open_ai_key, msg_txt, chat_bot, doc_text_state, doc_emb_state], [chat_bot]) # 发送消息
|
| 181 |
|
| 182 |
if __name__ == "__main__":
|