bioguider 0.2.14__tar.gz → 0.2.15__tar.gz

{bioguider-0.2.14 → bioguider-0.2.15}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: bioguider
-Version: 0.2.14
+Version: 0.2.15
 Summary: An AI-Powered package to help biomedical developers to generate clear documentation
 License: MIT
 Author: Cankun Wang

{bioguider-0.2.14 → bioguider-0.2.15}/bioguider/agents/agent_task.py

@@ -5,7 +5,7 @@ from abc import ABC, abstractmethod
 from langchain_openai.chat_models.base import BaseChatOpenAI
 from langgraph.graph.graph import CompiledGraph
 
-from bioguider.utils.constants import DEFAULT_TOKEN_USAGE
+from bioguider.utils.constants import DEFAULT_TOKEN_USAGE, MAX_STEP_COUNT
 from bioguider.database.summarized_file_db import SummarizedFilesDb
 
 class AgentTask(ABC):
@@ -82,7 +82,7 @@ class AgentTask(ABC):
         for s in self.graph.stream(
             input=input, 
             stream_mode="values",
-            config={"recursion_limit": 500},
+            config={"recursion_limit": MAX_STEP_COUNT},
         ):
             print(s)
 

{bioguider-0.2.14 → bioguider-0.2.15}/bioguider/agents/collection_observe_step.py

@@ -8,6 +8,7 @@ from bioguider.agents.collection_task_utils import CollectionWorkflowState
 from bioguider.agents.common_agent_2step import CommonAgentTwoChainSteps, CommonAgentTwoSteps
 from bioguider.agents.peo_common_step import PEOCommonStep
 from bioguider.agents.prompt_utils import COLLECTION_GOAL, COLLECTION_PROMPTS
+from bioguider.utils.constants import MAX_STEP_COUNT
 
 
 COLLECTION_OBSERVE_SYSTEM_PROMPT = """You are an expert software developer and technical documentation analyst.
@@ -92,11 +93,14 @@ class CollectionObserveStep(PEOCommonStep):
             important_instructions=important_instructions,
         )
     def _execute_directly(self, state: CollectionWorkflowState):
+        step_count = state["step_count"]
+        instruction = "Now, we have reached max recursion limit, please give me the **final answer** based on the current information" \
+            if step_count == MAX_STEP_COUNT - 2 else "Let's begin thinking."
         system_prompt = self._build_prompt(state)
         agent = CommonAgentTwoSteps(llm=self.llm)
         res, _, token_usage, reasoning_process = agent.go(
             system_prompt=system_prompt,
-            instruction_prompt="Let's begin thinking.",
+            instruction_prompt=instruction,
             schema=ObservationResult,
         )
         state["final_answer"] = res.FinalAnswer
@@ -104,9 +108,10 @@ class CollectionObserveStep(PEOCommonStep):
         thoughts = res.Thoughts
         state["step_analysis"] = analysis
         state["step_thoughts"] = thoughts
+        state["step_count"] += 1
         self._print_step(
             state,
-            step_output=f"**Observation Reasoning Process**\n{reasoning_process}"
+            step_output=f"**Observation Reasoning Process: {state['step_count']}**\n{reasoning_process}"
         )
         self._print_step(
             state,

{bioguider-0.2.14 → bioguider-0.2.15}/bioguider/agents/collection_task.py

@@ -176,7 +176,7 @@ class CollectionTask(AgentTask):
         self.graph = graph.compile()
 
     def collect(self) -> list[str] | None:
-        s = self._go_graph({"goal_item": self.goal_item})
+        s = self._go_graph({"goal_item": self.goal_item, "step_count": 0})
         if s is None or 'final_answer' not in s:
             return None
         if s["final_answer"] is None:

{bioguider-0.2.14 → bioguider-0.2.15}/bioguider/agents/collection_task_utils.py

@@ -28,6 +28,7 @@ class CollectionWorkflowState(TypedDict):
 
     goal_item: Optional[str]
     final_answer: Optional[str]
+    step_count: Optional[int]
 
 RELATED_FILE_GOAL_ITEM = """
 Your task is to determine whether the file is related to **{goal_item}**.

{bioguider-0.2.14 → bioguider-0.2.15}/bioguider/agents/identification_observe_step.py

@@ -5,6 +5,7 @@ from bioguider.agents.agent_utils import ObservationResult
 from bioguider.agents.common_agent_2step import CommonAgentTwoSteps, CommonAgentTwoChainSteps
 from bioguider.agents.identification_task_utils import IdentificationWorkflowState
 from bioguider.agents.peo_common_step import PEOWorkflowState, PEOCommonStep
+from bioguider.utils.constants import MAX_STEP_COUNT
 
 
 ## observation system prompt
@@ -76,11 +77,14 @@ class IdentificationObserveStep(PEOCommonStep):
         )
 
     def _execute_directly(self, state: IdentificationWorkflowState):
+        step_count = state["step_count"]
+        instruction = "Now, we have reached max recursion limit, please give me the **final answer** based on the current information" \
+            if step_count == MAX_STEP_COUNT - 2 else "Now, Let's begin."
         system_prompt = self._prepare_system_prompt(state)
         agent = CommonAgentTwoSteps(llm=self.llm)
         res, _, token_usage, reasoning_process = agent.go(
             system_prompt=system_prompt,
-            instruction_prompt="Now, let's begin.",
+            instruction_prompt=instruction,
             schema=ObservationResult,
         )
         state["final_answer"] = res.FinalAnswer
@@ -88,9 +92,10 @@ class IdentificationObserveStep(PEOCommonStep):
         thoughts = res.Thoughts
         state["step_analysis"] = analysis
         state["step_thoughts"] = thoughts
+        state["step_count"] += 1
         self._print_step(
             state,
-            step_output=f"**Observation Reasoning Process**\n{reasoning_process}"
+            step_output=f"**Observation Reasoning Process {state['step_count']}**\n{reasoning_process}"
         )
         self._print_step(
             state,

{bioguider-0.2.14 → bioguider-0.2.15}/bioguider/agents/identification_task.py

@@ -173,6 +173,7 @@ class IdentificationTask(AgentTask):
         s = self._go_graph({
             "goal": IDENTIFICATION_GOAL_PROJECT_TYPE,
             "final_answer_example": PROJECT_TYPE_FINAL_ANSWER_EXAMPLE,
+            "step_count": 0,
         })
         proj_type = s["final_answer"] if "final_answer" in s else "unknown type"
         return self._parse_project_type(proj_type)
@@ -181,6 +182,7 @@ class IdentificationTask(AgentTask):
         s = self._go_graph({
             "goal": IDENTIFICATION_GOAL_PRIMARY_LANGUAGE,
             "final_answer_example": PRIMARY_LANGUAGE_FINAL_ANSWER_EXAMPLE,
+            "step_count": 0,
         })
         language = s["final_answer"] if "final_answer" in s else "unknown type"
         return self._parse_primary_language(language)
@@ -189,6 +191,7 @@ class IdentificationTask(AgentTask):
         s = self._go_graph({
             "goal": IDENTIFICATION_GOAL_META_DATA,
             "final_answer_example": META_DATA_FINAL_ANSWER_EXAMPLE,
+            "step_count": 0,
         })
         meta_data = s["final_answer"] if "final_answer" in s else "unknown type"
         return self._parse_meta_data(meta_data)
@@ -205,6 +208,7 @@ class IdentificationTask(AgentTask):
             "final_answer_example": final_answer_example,
             "plan_instructions": plan_instructions,
             "observe_instructions": observe_instructions,
+            "step_count": 0,
         })
         return s["final_answer"] if "final_answer" in s else None
 

{bioguider-0.2.14 → bioguider-0.2.15}/bioguider/agents/identification_task_utils.py

@@ -18,3 +18,5 @@ class IdentificationWorkflowState(TypedDict):
     step_output: Optional[str]
     step_analysis: Optional[str]
     step_thoughts: Optional[str]
+
+    step_count: Optional[int]

{bioguider-0.2.14 → bioguider-0.2.15}/bioguider/agents/prompt_utils.py

@@ -208,6 +208,7 @@ If **any one** of these is present, the document should be classified as Contrib
 """,
         "plan_important_instructions": """ - A comiled standalone software file is non-textual and appears to be in an executable format (e.g., `.exe`, `.dll`, `.so`, `.bin`, `.elf`).
  - A comiled standalone software file **is not a script or compiled library**, that is, It is not a wrapper script (e.g., shell, Python, Python notebook or Rmd) nor a dynamic/shared library meant for linking.
+   So, when you are identifying a binary file, **do not** use any tools (our tools don't work for binary file), you need to figure out if it is compiled standalone software file by the file name and extension on your own.
  - When identifying source code file, prioritize analyzing the file's **extension** and **file name** and try to avoid reading file, using check_file_related_tool or summarizing file content.
  - When identifying example data, prioritize analyzing the file's **extension** (like .dat, .csv, .fastq, and so on) and **file name** (like example_data.txt, example.dat, and so on). If extension/name is ambiguous, use summarizing file content to decide.
  - **Note**: You **only need to detect** whether at least **one** compiled standalone software file, **one** source code file and **one** example data file exist — no need to list all such files.

{bioguider-0.2.14 → bioguider-0.2.15}/bioguider/utils/constants.py

@@ -39,4 +39,5 @@ class ProjectMetadata:
         self.license = license
 
 MAX_FILE_LENGTH=10 *1024 # 10K
-MAX_SENTENCE_NUM=20
+MAX_SENTENCE_NUM=20
+MAX_STEP_COUNT=30

{bioguider-0.2.14 → bioguider-0.2.15}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "bioguider"
-version = "0.2.14"
+version = "0.2.15"
 description = "An AI-Powered package to help biomedical developers to generate clear documentation"
 authors = [
     "Cankun Wang <Cankun.Wang@osumc.edu>",