bioguider 0.2.9__tar.gz → 0.2.10__tar.gz

{bioguider-0.2.9 → bioguider-0.2.10}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: bioguider
-Version: 0.2.9
+Version: 0.2.10
 Summary: An AI-Powered package to help biomedical developers to generate clear documentation
 License: MIT
 Author: Cankun Wang

{bioguider-0.2.9 → bioguider-0.2.10}/bioguider/agents/common_agent_2step.py

@@ -154,7 +154,7 @@ class CommonAgentTwoChainSteps(CommonAgentTwoSteps):
     def _invoke_agent(self, system_prompt, instruction_prompt, schema, post_process = None, **kwargs):
         # Initialize the callback handler
         callback_handler = OpenAICallbackHandler()
-        processed_system_prompt = system_prompt.replace("{", "(").replace("}", ")")
+        processed_system_prompt = system_prompt.replace("{", "{{").replace("}", "}}")
         cot_prompt = self._build_prompt_for_cot_step(
             system_prompt=processed_system_prompt, 
             instruction_prompt=instruction_prompt

{bioguider-0.2.9 → bioguider-0.2.10}/bioguider/agents/evaluation_task.py

@@ -200,7 +200,7 @@ class EvaluationTask(ABC):
 
 class EvaluationREADMEResult(BaseModel):
     project_level: Optional[bool]=Field(description="A boolean value specifying if the README file is **project-level** README. TRUE: project-level, FALSE, folder-level")
-    score: Optional[float]=Field(description="An overall score")
+    score: Optional[str]=Field(description="An overall score")
     key_strengths: Optional[str]=Field(description="A string specifying the key strengths of README file.")
     overall_improvement_suggestions: Optional[list[str]]=Field(description="A list of overall improvement suggestions")
 
@@ -214,7 +214,7 @@ EvaluationREADMEResultSchema = {
             "title": "Project Level"
         },
         "score": {
-            "anyOf": [{"type": "number"}, {"type": "null"}],
+            "anyOf": [{"type": "string"}, {"type": "null"}],
             "description": "An overall score",
             "title": "Score"
         },
@@ -260,7 +260,7 @@ class EvaluationREADMETask(EvaluationTask):
                 readme_evaluations[readme_file] = {
                     "evaluation": {
                         "project_level": "/" in readme_file,
-                        "score": 0,
+                        "score": "Poor",
                         "key_strengths": f"{readme_file} is an empty file.",
                         "overall_improvement_suggestions": f"{readme_file} is an empty file.",
                     },

{bioguider-0.2.9 → bioguider-0.2.10}/bioguider/agents/identification_observe_step.py

@@ -25,7 +25,8 @@ Carefully review the **Goal**, **Repository File Structure**, and **Intermediate
   - Then provide your result under **FinalAnswer**  
   ```
   **Analysis**: your analysis here  
-  **FinalAnswer**: your final answer here, in json format **without** json fence (```json ... ```), like {final_answer_example}
+  **FinalAnswer**: your final answer here, in **raw json format**, **including** the surrounding "{{}}" but **without** using code fence (```json ... ```), 
+  For example, output exactly: {final_answer_example}
   ```
 - If the information is **not sufficient** to achieve the goal, simply explain why under **Thoughts**:  
   ```

{bioguider-0.2.9 → bioguider-0.2.10}/bioguider/managers/evaluation_manager.py

@@ -25,8 +25,8 @@ class EvaluationManager:
         self.repo_url = repo_url
         self.rag = RAG()
         self.rag.initialize_db_manager()
-        self.rag.prepare_retriever(repo_url_or_path=repo_url)
-
+        self.rag.initialize_repo(repo_url_or_path=repo_url)
+        
         author, repo_name = parse_repo_url(repo_url)
         self.summary_file_db = SummarizedFilesDb(author, repo_name)
 

{bioguider-0.2.9 → bioguider-0.2.10}/bioguider/rag/data_pipeline.py

@@ -504,7 +504,11 @@ class DatabaseManager:
         self.repo_url_or_path = None
         self.repo_paths = None
 
-    def prepare_database(self, repo_url_or_path: str, access_token: str = None) -> Tuple[List[Document], List[Document]]:
+    def reset_database_and_create_repo(self, repo_url_or_path: str, access_token: str = None):
+        self._reset_database()
+        self._create_repo(repo_url_or_path, access_token)
+
+    def prepare_database(self) -> Tuple[List[Document], List[Document]]:
         """
         Create a new database from the repository.
 
@@ -515,9 +519,7 @@ class DatabaseManager:
         Returns:
             Tuple[List[Document], List[Document]]: Tuple of two Lists of Document objects
         """
-        self.reset_database()
-        self._create_repo(repo_url_or_path, access_token)
-        return self.prepare_db_index()
+        return self._prepare_db_index()
     
     def _extract_repo_name_from_url(self, repo_url_or_path: str, repo_type: str) -> str:
         # Extract owner and repo name to create unique identifier
@@ -534,7 +536,7 @@ class DatabaseManager:
             repo_name = url_parts[-1].replace(".git", "")
         return repo_name
 
-    def reset_database(self):
+    def _reset_database(self):
         """
         Reset the database to its initial state.
         """
@@ -608,7 +610,7 @@ class DatabaseManager:
             return self.repo_paths["save_repo_dir"]
         return None
     
-    def prepare_db_index(self) -> Tuple[List[Document], List[Document]]:
+    def _prepare_db_index(self) -> Tuple[List[Document], List[Document]]:
         """
         Prepare the indexed database for the repository.
         :return: Tuple of two Lists of Document objects
@@ -647,16 +649,3 @@ class DatabaseManager:
         logger.info(f"Total transformed code documents: {len(transformed_code_documents)}")
         return transformed_doc_documents, transformed_code_documents
 
-    def prepare_retriever(self, repo_url_or_path: str, access_token: str = None):
-        """
-        Prepare the retriever for a repository.
-        This is a compatibility method for the isolated API.
-
-        Args:
-            repo_url_or_path (str): The URL or local path of the repository
-            access_token (str, optional): Access token for private repositories
-
-        Returns:
-            List[Document]: List of Document objects
-        """
-        return self.prepare_database(repo_url_or_path, access_token)

{bioguider-0.2.9 → bioguider-0.2.10}/bioguider/rag/rag.py

@@ -50,22 +50,25 @@ class RAG(adal.Component):
     def initialize_db_manager(self):
         """Initialize the database manager with local storage"""
         self.db_manager = DatabaseManager()
-        self.transformed_doc_documents = []
-        self.transformed_code_documents = []
+        self.transformed_doc_documents: list | None = None
+        self.transformed_code_documents: list | None = None
+        self.access_token: str | None = None
 
-    def prepare_retriever(self, repo_url_or_path: str, access_token: str = None):
+    def initialize_repo(self, repo_url_or_path: str, access_token: str = None):
+        self.repo_url_or_path = repo_url_or_path
+        self.access_token = access_token
+        self.db_manager.reset_database_and_create_repo(repo_url_or_path, access_token)
+
+    def _prepare_retriever(self):
         """
         Prepare the retriever for a repository.
         Will load database from local storage if available.
-
-        Args:
-            repo_url_or_path: URL or local path to the repository
-            access_token: Optional access token for private repositories
         """
-        self.initialize_db_manager()
-        self.repo_url_or_path = repo_url_or_path
+        if self.transformed_code_documents is not None and self.transformed_doc_documents is not None:
+            # retrievers have been prepared
+            return
         self.transformed_doc_documents, self.transformed_code_documents \
-            = self.db_manager.prepare_database(repo_url_or_path, access_token)
+            = self.db_manager.prepare_database()
         logger.info(f"Loaded {len(self.transformed_doc_documents)} doc documents for retrieval")
         logger.info(f"Loaded {len(self.transformed_code_documents)} code documents for retrieval")
         self.doc_retriever = FAISSRetriever(
@@ -93,6 +96,7 @@ class RAG(adal.Component):
         Returns:
             retrieved_documents: List of documents retrieved based on the query
         """
+        self._prepare_retriever()
         retrieved_documents = self.doc_retriever(query)
         # Fill in the documents
         retrieved_documents[0].documents = [

{bioguider-0.2.9 → bioguider-0.2.10}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "bioguider"
-version = "0.2.9"
+version = "0.2.10"
 description = "An AI-Powered package to help biomedical developers to generate clear documentation"
 authors = [
     "Cankun Wang <Cankun.Wang@osumc.edu>",