Add list <-> tagged text utilities to HTB class

src/backend/gradio_highlightedtextbox/highlightedtextbox.py

@@ -1,14 +1,14 @@
 from __future__ import annotations
 
-from typing import Any, Callable, List, Tuple
+from typing import Any, Callable
 
-from gradio.data_classes import GradioRootModel
 from gradio.components.base import FormComponent
+from gradio.data_classes import GradioRootModel
 from gradio.events import Events
 
 
 class HighlightedTextData(GradioRootModel):
-    root: List[Tuple[str, str | None]]
+    root: list[tuple[str, str | None]]
 
 
 class HighlightedTextbox(FormComponent):
@@ -16,12 +16,12 @@ class HighlightedTextbox(FormComponent):
     Creates a textarea for user to enter string input or display string output where some
     elements are highlighted.
     Preprocessing: passes a list of tuples as a {List[Tuple[str, float | str | None]]]} into the function. If no labels are provided, the text will be displayed as a single span.
-    Postprocessing: expects a {List[Tuple[str, float | str]]]} consisting of spans of text and their associated labels, or a {Dict} with two keys: 
-        (1) "text" whose value is the complete text, and 
-        (2) "highlights", which is a list of dictionaries, each of which have the keys: 
-            "highlight_type" (consisting of the highlight label), 
-            "start" (the character index where the label starts), and 
-            "end" (the character index where the label ends). 
+    Postprocessing: expects a {List[Tuple[str, float | str]]]} consisting of spans of text and their associated labels, or a {Dict} with two keys:
+        (1) "text" whose value is the complete text, and
+        (2) "highlights", which is a list of dictionaries, each of which have the keys:
+            "highlight_type" (consisting of the highlight label),
+            "start" (the character index where the label starts), and
+            "end" (the character index where the label ends).
         Highlights should not overlap.
     """
 
@@ -114,15 +114,15 @@ class HighlightedTextbox(FormComponent):
             every=every,
         )
 
-    def preprocess(self, payload: str | None) -> str | None:
-        return None if payload is None else str(payload)
+    def preprocess(self, payload: HighlightedTextData) -> dict:
+        return {"id": self.elem_id, "data": payload.root}
 
     def postprocess(
-        self, y: HighlightedTextData | dict | None
-    ) -> HighlightedTextData | None:
+        self, y: list[tuple[str, str | None]] | dict | None
+    ) -> list[tuple[str, str | None]] | None:
         """
         Parameters:
-            y: List of (word, category) tuples, or a dictionary of two keys: "text", and "highlights", which itself is 
+            y: List of (word, category) tuples, or a dictionary of two keys: "text", and "highlights", which itself is
                 a list of dictionaries, each of which have the keys: "highlight_type", "start", and "end"
         Returns:
             List of (word, category) tuples
@@ -178,3 +178,94 @@ class HighlightedTextbox(FormComponent):
 
     def example_inputs(self) -> Any:
         return [("Hello", None), ("world", "highlight")]
+
+    @classmethod
+    def tagged_text_to_tuples(
+        cls, text: str, tag_id: str, tag_open: str = "<h>", tag_close: str = "</h>"
+    ) -> list[tuple[str, str | None]]:
+        """Parse a text containing tags into a list of tuples in the format accepted by HighlightedTextbox.
+
+        E.g. Hello <h>world</h>! -> [("Hello", None), ("world", <TAG_ID>), ("!", None)]
+
+        Args:
+            text (`str`):
+                Text containing tags that needs to be parsed.
+            tag_id (`str`):
+                Label to use for the second element of the tuple.
+            tag_open (`str`, *optional*, defaults to "<h>"):
+                Tag used to mark the beginning of a highlighted section.
+            tag_close (`str`, *optional*, defaults to "</h>"):
+                Tag used to mark the end of a highlighted section.
+
+        Raises:
+            `ValueError`: Number of open tags does not match number of closed tags.
+
+        Returns:
+            `list[tuple[str, str | None]]`: List of tuples in the format accepted by HighlightedTextbox.
+        """
+        # Check that the text is well-formed (i.e. no nested or empty tags)
+        num_tags = text.count(tag_open)
+        if num_tags != text.count(tag_close):
+            raise ValueError(
+                f"Number of open tags ({tag_open}) does not match number of closed tags ({tag_close})."
+            )
+        elif num_tags == 0:
+            return [(text, None)]
+        elif num_tags > 0:
+            out = []
+            pre_tag_text = text[: text.index(tag_open)]
+            if pre_tag_text:
+                out += [(pre_tag_text.strip(), None)]
+
+            tag_text = text[
+                text.index(tag_open) + len(tag_open) : text.index(tag_close)
+            ]
+            out += [(tag_text.strip(), tag_id)]
+            if num_tags > 1:
+                remaining_text = text[text.index(tag_close) + len(tag_close) :]
+                out += cls.tagged_text_to_tuples(
+                    remaining_text,
+                    tag_id=tag_id,
+                    tag_open=tag_open,
+                    tag_close=tag_close,
+                )
+            else:
+                post_tag_text = text[text.index(tag_close) + len(tag_close) :]
+                if post_tag_text:
+                    out += [(post_tag_text, None)]
+        return out
+
+    @staticmethod
+    def tuples_to_tagged_text(
+        tuples: list[tuple[str, str | None]],
+        tag_ids: str | list[str] = [],
+        tag_open: str = "<h>",
+        tag_close: str = "</h>",
+    ) -> str:
+        """Convert a list of tuples in the format accepted by HighlightedTextbox into a text containing tags.
+
+        E.g. [("Hello", None), ("world", <TAG_ID>), ("!", None)] -> Hello <h>world</h>!
+
+        Args:
+            tuples (`list[tuple[str, str | None]]`):
+                List of tuples in the format accepted by HighlightedTextbox.
+            tag_ids (`str` | `list[str]`):
+                Label(s) to select for the second element of the tuple. All other labels will be ignored
+                (i.e. replaced with None)
+            tag_open (`str`, *optional*, defaults to "<h>"):
+                Tag used to mark the beginning of a highlighted section.
+            tag_close (`str`, *optional*, defaults to "</h>"):
+                Tag used to mark the end of a highlighted section.
+
+        Returns:
+            `str`: Text containing tags.
+        """
+        if isinstance(tag_ids, str):
+            tag_ids = [tag_ids]
+        out = ""
+        for text, tag_id in tuples:
+            if tag_id in tag_ids:
+                out += f" {tag_open}{text.strip()}{tag_close} "
+            else:
+                out += text.strip()
+        return out.strip()

src/backend/gradio_highlightedtextbox/highlightedtextbox.pyi

@@ -12,12 +12,12 @@ class HighlightedTextbox(FormComponent):
     Creates a textarea for user to enter string input or display string output where some
     elements are highlighted.
     Preprocessing: passes a list of tuples as a {List[Tuple[str, float | str | None]]]} into the function. If no labels are provided, the text will be displayed as a single span.
-    Postprocessing: expects a {List[Tuple[str, float | str]]]} consisting of spans of text and their associated labels, or a {Dict} with two keys: 
-        (1) "text" whose value is the complete text, and 
-        (2) "highlights", which is a list of dictionaries, each of which have the keys: 
-            "highlight_type" (consisting of the highlight label), 
-            "start" (the character index where the label starts), and 
-            "end" (the character index where the label ends). 
+    Postprocessing: expects a {List[Tuple[str, float | str]]]} consisting of spans of text and their associated labels, or a {Dict} with two keys:
+        (1) "text" whose value is the complete text, and
+        (2) "highlights", which is a list of dictionaries, each of which have the keys:
+            "highlight_type" (consisting of the highlight label),
+            "start" (the character index where the label starts), and
+            "end" (the character index where the label ends).
         Highlights should not overlap.
     """
 
@@ -110,15 +110,15 @@ class HighlightedTextbox(FormComponent):
             every=every,
         )
 
-    def preprocess(self, payload: str | None) -> str | None:
-        return None if payload is None else str(payload)
+    def preprocess(self, payload: HighlightedTextData) -> dict:
+        return {"id": self.elem_id, "data": payload.root}
 
     def postprocess(
-        self, y: HighlightedTextData | dict | None
-    ) -> HighlightedTextData | None:
+        self, y: list[tuple[str, str | None]] | dict | None
+    ) -> list[tuple[str, str | None]] | None:
         """
         Parameters:
-            y: List of (word, category) tuples, or a dictionary of two keys: "text", and "highlights", which itself is 
+            y: List of (word, category) tuples, or a dictionary of two keys: "text", and "highlights", which itself is
                 a list of dictionaries, each of which have the keys: "highlight_type", "start", and "end"
         Returns:
             List of (word, category) tuples
@@ -175,6 +175,97 @@ class HighlightedTextbox(FormComponent):
     def example_inputs(self) -> Any:
         return [("Hello", None), ("world", "highlight")]
 
+    @classmethod
+    def tagged_text_to_tuples(
+        cls, text: str, tag_id: str, tag_open: str = "<h>", tag_close: str = "</h>"
+    ) -> list[tuple[str, str | None]]:
+        """Parse a text containing tags into a list of tuples in the format accepted by HighlightedTextbox.
+
+        E.g. Hello <h>world</h>! -> [("Hello", None), ("world", <TAG_ID>), ("!", None)]
+
+        Args:
+            text (`str`):
+                Text containing tags that needs to be parsed.
+            tag_id (`str`):
+                Label to use for the second element of the tuple.
+            tag_open (`str`, *optional*, defaults to "<h>"):
+                Tag used to mark the beginning of a highlighted section.
+            tag_close (`str`, *optional*, defaults to "</h>"):
+                Tag used to mark the end of a highlighted section.
+
+        Raises:
+            `ValueError`: Number of open tags does not match number of closed tags.
+
+        Returns:
+            `list[tuple[str, str | None]]`: List of tuples in the format accepted by HighlightedTextbox.
+        """
+        # Check that the text is well-formed (i.e. no nested or empty tags)
+        num_tags = text.count(tag_open)
+        if num_tags != text.count(tag_close):
+            raise ValueError(
+                f"Number of open tags ({tag_open}) does not match number of closed tags ({tag_close})."
+            )
+        elif num_tags == 0:
+            return [(text, None)]
+        elif num_tags > 0:
+            out = []
+            pre_tag_text = text[: text.index(tag_open)]
+            if pre_tag_text:
+                out += [(pre_tag_text.strip(), None)]
+
+            tag_text = text[
+                text.index(tag_open) + len(tag_open) : text.index(tag_close)
+            ]
+            out += [(tag_text.strip(), tag_id)]
+            if num_tags > 1:
+                remaining_text = text[text.index(tag_close) + len(tag_close) :]
+                out += cls.tagged_text_to_tuples(
+                    remaining_text,
+                    tag_id=tag_id,
+                    tag_open=tag_open,
+                    tag_close=tag_close,
+                )
+            else:
+                post_tag_text = text[text.index(tag_close) + len(tag_close) :]
+                if post_tag_text:
+                    out += [(post_tag_text, None)]
+        return out
+
+    @staticmethod
+    def tuples_to_tagged_text(
+        tuples: list[tuple[str, str | None]],
+        tag_ids: str | list[str] = [],
+        tag_open: str = "<h>",
+        tag_close: str = "</h>",
+    ) -> str:
+        """Convert a list of tuples in the format accepted by HighlightedTextbox into a text containing tags.
+
+        E.g. [("Hello", None), ("world", <TAG_ID>), ("!", None)] -> Hello <h>world</h>!
+
+        Args:
+            tuples (`list[tuple[str, str | None]]`):
+                List of tuples in the format accepted by HighlightedTextbox.
+            tag_ids (`str` | `list[str]`):
+                Label(s) to select for the second element of the tuple. All other labels will be ignored
+                (i.e. replaced with None)
+            tag_open (`str`, *optional*, defaults to "<h>"):
+                Tag used to mark the beginning of a highlighted section.
+            tag_close (`str`, *optional*, defaults to "</h>"):
+                Tag used to mark the end of a highlighted section.
+
+        Returns:
+            `str`: Text containing tags.
+        """
+        if isinstance(tag_ids, str):
+            tag_ids = [tag_ids]
+        out = ""
+        for text, tag_id in tuples:
+            if tag_id in tag_ids:
+                out += f" {tag_open}{text.strip()}{tag_close} "
+            else:
+                out += text.strip()
+        return out.strip()
+
     
     def change(self,
         fn: Callable | None,

src/demo/app.py

@@ -1,20 +1,66 @@
 import gradio as gr
 from gradio_highlightedtextbox import HighlightedTextbox
 
-def set_highlighted():
-    return HighlightedTextbox(
-        value=[("Non è qualcosa di cui vergognarsi: non è diverso dalle paure e", None), ("odie", "Potential issue"), ("personali", None), ("di altre cose", "Potential issue"), ("che", None), ("molta gente ha", "Potential issue"), (".", None)],
-        interactive=True, label="Output", show_legend=True, show_label=False, legend_label="Test:", show_legend_label=True
+
+def convert_tagged_text_to_highlighted_text(
+    tagged_text: str, tag_id: str, tag_open: str, tag_close: str
+) -> list[tuple[str, str | None]]:
+    return HighlightedTextbox.tagged_text_to_tuples(
+        tagged_text, tag_id, tag_open, tag_close
     )
 
+
 with gr.Blocks() as demo:
+    tag_id = gr.Textbox(
+        "Potential issue",
+        label="Tag ID",
+        show_label=True,
+        info="Insert a tag ID to use in the highlighted textbox.",
+    )
+    tag_open = gr.Textbox(
+        "<h>",
+        label="Tag open",
+        show_label=True,
+        info="Insert a tag to mark the beginning of a highlighted section.",
+    )
+    tag_close = gr.Textbox(
+        "</h>",
+        label="Tag close",
+        show_label=True,
+        info="Insert a tag to mark the end of a highlighted section.",
+    )
     with gr.Row():
-        gr.Textbox(" It is not something to be ashamed of: it is no different from the personal fears and dislikes of other things that very many people have.", interactive=False)
+        tagged = gr.Textbox(
+            "It is not something to be ashamed of: it is no different from the <h>personal fears</h> and <h>dislikes</h> of other things that <h>very many people</h> have.",
+            interactive=True,
+            label="Input",
+            show_label=True,
+            info="Insert a text with <h>...</h> tags to mark spans that will be highlighted.",
+        )
         high = HighlightedTextbox(
-            interactive=True, label="Input", show_legend=True, show_label=False, legend_label="Legend:", show_legend_label=True
+            interactive=True,
+            label="Output",
+            info="Highlighted textbox.",
+            show_legend=True,
+            show_label=True,
+            legend_label="Legend:",
+            show_legend_label=True,
         )
     button = gr.Button("Submit")
-    button.click(fn=set_highlighted, inputs=[], outputs=high)
-    
+    button.click(
+        fn=convert_tagged_text_to_highlighted_text,
+        inputs=[tagged, tag_id, tag_open, tag_close],
+        outputs=high,
+    )
+    # Initialization does not work
+    high = HighlightedTextbox(
+        convert_tagged_text_to_highlighted_text(
+            tagged.value, tag_id.value, tag_open.value, tag_close.value
+        ),
+        interactive=True,
+        label="Does not work",
+        show_label=True,
+    )
+
 
-demo.launch()
+demo.launch()