fix(plan, rag) fix the bug in the word reader and plan notebook class.

src/agentscope/plan/_in_memory_storage.py

@@ -14,6 +14,15 @@ def __init__(self) -> None:
         super().__init__()
         self.plans = OrderedDict()
 
+        # Support historical plan serialization and deserialization
+        self.register_state(
+            "plans",
+            lambda plans: {k: v.model_dump() for k, v in plans.items()},
+            lambda json_data: OrderedDict(
+                (k, Plan.model_validate(v)) for k, v in json_data.items()
+            ),
+        )
+
     async def add_plan(self, plan: Plan, override: bool = True) -> None:
         """Add a plan to the storage.
 

src/agentscope/plan/_plan_model.py

@@ -33,7 +33,6 @@ class SubTask(BaseModel):
     )
     outcome: str | None = Field(
         description="The actual outcome of the subtask.",
-        exclude=True,
         default=None,
     )
     state: Literal["todo", "in_progress", "done", "abandoned"] = Field(
@@ -48,7 +47,6 @@ class SubTask(BaseModel):
     finished_at: str | None = Field(
         description="The time the subtask was finished.",
         default=None,
-        exclude=True,
     )
 
     def finish(self, outcome: str) -> None:
@@ -106,7 +104,7 @@ def to_markdown(self, detailed: bool = False) -> str:
 class Plan(BaseModel):
     """The plan model used in the plan module, contains a list of subtasks."""
 
-    id: str = Field(exclude=True, default_factory=shortuuid.uuid)
+    id: str = Field(default_factory=shortuuid.uuid)
     name: str = Field(
         description=(
             "The plan name, should be concise, descriptive and not exceed 10 "
@@ -133,24 +131,41 @@ class Plan(BaseModel):
     created_at: str = Field(
         description="The time the plan was created.",
         default_factory=_get_timestamp,
-        exclude=True,
     )
     state: Literal["todo", "in_progress", "done", "abandoned"] = Field(
         description="The state of the plan.",
         default="todo",
-        exclude=True,
     )
     finished_at: str | None = Field(
         description="The time the plan was finished.",
         default=None,
-        exclude=True,
     )
     outcome: str | None = Field(
         description="The actual outcome of the plan.",
         default=None,
-        exclude=True,
     )
 
+    def refresh_plan_state(self) -> str:
+        """Refresh the plan state based on the states of its subtasks. This
+        function only switches the plan state between "todo" and "in_progress".
+
+        # TODO: Handle the plan state much more formally.
+        """
+        if self.state in ["done", "abandoned"]:
+            return ""
+
+        any_in_progress = any(_.state == "in_progress" for _ in self.subtasks)
+
+        if any_in_progress and self.state == "todo":
+            self.state = "in_progress"
+            return "The plan state has been updated to 'in_progress'."
+
+        elif not any_in_progress and self.state == "in_progress":
+            self.state = "todo"
+            return "The plan state has been updated to 'todo'."
+
+        return ""
+
     def finish(
         self,
         state: Literal["done", "abandoned"],

src/agentscope/plan/_plan_notebook.py

@@ -424,7 +424,7 @@ async def revise_current_plan(
     async def update_subtask_state(
         self,
         subtask_idx: int,
-        state: Literal["todo", "in_progress", "deprecated"],
+        state: Literal["todo", "in_progress", "abandoned"],
     ) -> ToolResponse:
         """Update the state of a subtask by given index and state. Note if you
         want to mark a subtask as done, you SHOULD call `finish_subtask`
@@ -487,7 +487,7 @@ async def update_subtask_state(
                 # Check all previous subtasks are done or deprecated
                 if idx < subtask_idx and subtask.state not in [
                     "done",
-                    "deprecated",
+                    "abandoned",
                 ]:
                     return ToolResponse(
                         content=[
@@ -520,14 +520,18 @@ async def update_subtask_state(
                     )
 
         self.current_plan.subtasks[subtask_idx].state = state
+
+        # Update the plan state to in_progress if not already
+        suffix = self.current_plan.refresh_plan_state()
+
         await self._trigger_plan_change_hooks()
         return ToolResponse(
             content=[
                 TextBlock(
                     type="text",
                     text=f"Subtask at index {subtask_idx}, named "
                     f"'{self.current_plan.subtasks[subtask_idx].name}' "
-                    f"is marked as '{state}' successfully.",
+                    f"is marked as '{state}' successfully. " + suffix,
                 ),
             ],
         )
@@ -585,7 +589,7 @@ async def finish_subtask(
         for idx, subtask in enumerate(
             self.current_plan.subtasks[0:subtask_idx],
         ):
-            if subtask.state not in ["done", "deprecated"]:
+            if subtask.state not in ["done", "abandoned"]:
                 return ToolResponse(
                     content=[
                         TextBlock(
@@ -815,12 +819,15 @@ def list_tools(
                 the agent.
         """
         return [
+            # subtask related tools
             self.view_subtasks,
             self.update_subtask_state,
             self.finish_subtask,
+            # plan related tools
             self.create_plan,
             self.revise_current_plan,
             self.finish_plan,
+            # historical plan related tools
             self.view_historical_plans,
             self.recover_historical_plan,
         ]

src/agentscope/rag/_reader/_word_reader.py

@@ -4,33 +4,41 @@
 import base64
 import hashlib
 import json
-from typing import Literal
+from typing import Literal, TYPE_CHECKING
 
-from docx.oxml import CT_P, CT_Tbl
-from docx.table import Table
-from docx.text.paragraph import Paragraph
-from docx.oxml.ns import qn
 
 from ._reader_base import ReaderBase
 from ._text_reader import TextReader
 from .._document import Document, DocMetadata
 from ..._logging import logger
 from ...message import ImageBlock, Base64Source, TextBlock
 
+if TYPE_CHECKING:
+    from docx.table import Table as DocxTable
+    from docx.text.paragraph import Paragraph as DocxParagraph
+else:
+    DocxTable = "docx.table.Table"
+    DocxParagraph = "docx.text.paragraph.Paragraph"
 
-def _extract_text_from_paragraph(para: Paragraph) -> str:
+
+def _extract_text_from_paragraph(para: DocxParagraph) -> str:
     """Extract text from a paragraph, including text in text boxes and shapes.
 
     Args:
-        para: Paragraph object
+        para (`Paragraph`):
+            The paragraph object from which to extract text.
+
 
     Returns:
-        str: Extracted text
+        `str`:
+            Extracted text
     """
     text = ""
 
     # Method 1: Extract all w:t elements directly from XML
     #  (handles revisions, hyperlinks, etc.)
+    from docx.oxml.ns import qn
+
     for t_elem in para._element.findall(".//" + qn("w:t")):
         if t_elem.text:
             text += t_elem.text
@@ -62,7 +70,7 @@ def _extract_text_from_paragraph(para: Paragraph) -> str:
     return text.strip()
 
 
-def _extract_table_data(table: Table) -> list[list[str]]:
+def _extract_table_data(table: DocxTable) -> list[list[str]]:
     """Extract table data, handling merged cells and preserving line breaks
     within cells.
 
@@ -75,6 +83,8 @@ def _extract_table_data(table: Table) -> list[list[str]]:
             Table data represented as a 2D list.
     """
 
+    from docx.oxml.ns import qn
+
     table_data = []
     # Extract table cell elements directly from XML
     for tr in table._element.findall(qn("w:tr")):
@@ -105,7 +115,7 @@ def _extract_table_data(table: Table) -> list[list[str]]:
     return table_data
 
 
-def _extract_image_data(para: Paragraph) -> list[ImageBlock]:
+def _extract_image_data(para: DocxParagraph) -> list[ImageBlock]:
     """Extract image data from a paragraph.
 
     Args:
@@ -118,6 +128,8 @@ def _extract_image_data(para: Paragraph) -> list[ImageBlock]:
     """
     images = []
 
+    from docx.oxml.ns import qn
+
     # Method 1: Find all drawing elements (modern Word format)
     drawings = para._element.findall(".//" + qn("w:drawing"))
 
@@ -334,6 +346,11 @@ def _get_data_blocks(self, word_path: str) -> list[TextBlock | ImageBlock]:
         # Read the Word document
         try:
             from docx import Document as DocxDocument
+            from docx.oxml import CT_P, CT_Tbl
+            from docx.text.paragraph import Paragraph
+            from docx.table import Table
+            from docx.oxml.ns import qn
+
         except ImportError as e:
             raise ImportError(
                 "Please install python-docx to use the Word reader. "