lean-explore 0.1.3__py3-none-any.whl → 0.2.0__py3-none-any.whl

lean_explore/local/service.py

@@ -52,10 +52,12 @@ class Service:
         default_faiss_k (int): Default number of FAISS neighbors to retrieve.
         default_pagerank_weight (float): Default weight for PageRank.
         default_text_relevance_weight (float): Default weight for text relevance.
-        default_name_match_weight (float): Default weight for name matching.
+        default_name_match_weight (float): Default weight for name matching (BM25).
         default_semantic_similarity_threshold (float): Default similarity threshold.
         default_results_limit (int): Default limit for search results.
         default_faiss_nprobe (int): Default nprobe for FAISS IVF indexes.
+        default_faiss_oversampling_factor (int): Default oversampling factor for
+            FAISS when package filters are active.
     """
 
     def __init__(self):
@@ -135,11 +137,8 @@ class Service:
         try:
             self.engine = create_engine(db_url)
             # Test connection
-            with (
-                self.engine.connect()
-            ):  # Ensure connect is within try for OperationalError
+            with self.engine.connect():  # type: ignore[attr-defined] # sqlalchemy stubs might be incomplete
                 logger.info("Database connection successful.")
-            # Setup SessionLocal after successful connection test
             self.SessionLocal: sessionmaker[SQLAlchemySessionType] = sessionmaker(
                 autocommit=False, autoflush=False, bind=self.engine
             )
@@ -147,7 +146,7 @@ class Service:
             guidance = (
                 "Please check your database configuration or connection parameters."
             )
-            if is_file_db:  # This check is now valid as is_file_db is defined earlier
+            if is_file_db:
                 guidance = (
                     f"The database file at '{db_path}' might be corrupted, "
                     "inaccessible, or not a valid SQLite file. "
@@ -180,6 +179,9 @@ class Service:
         )
         self.default_results_limit: int = defaults.DEFAULT_RESULTS_LIMIT
         self.default_faiss_nprobe: int = defaults.DEFAULT_FAISS_NPROBE
+        self.default_faiss_oversampling_factor: int = (
+            defaults.DEFAULT_FAISS_OVERSAMPLING_FACTOR
+        )
 
         logger.info("Local Service initialized successfully.")
 
@@ -257,18 +259,18 @@ class Service:
                     pagerank_weight=self.default_pagerank_weight,
                     text_relevance_weight=self.default_text_relevance_weight,
                     name_match_weight=self.default_name_match_weight,
+                    log_searches=True,
                     selected_packages=package_filters,
                     semantic_similarity_threshold=(
                         self.default_semantic_similarity_threshold
                     ),
                     faiss_nprobe=self.default_faiss_nprobe,
+                    faiss_oversampling_factor=self.default_faiss_oversampling_factor,
                 )
-            except Exception as e:  # Catch exceptions from perform_search
+            except Exception as e:
                 logger.error(
                     f"Error during perform_search execution: {e}", exc_info=True
                 )
-                # Re-raise to allow higher-level error handling if needed by the caller
-                # (e.g., MCP server might want to return a specific error response)
                 raise
 
         api_results = [
@@ -285,7 +287,7 @@ class Service:
             packages_applied=package_filters,
             results=final_results,
             count=len(final_results),
-            total_candidates_considered=len(api_results),  # Number before final limit
+            total_candidates_considered=len(api_results),
             processing_time_ms=processing_time_ms,
         )
 
@@ -314,10 +316,8 @@ class Service:
                     f"Database error in get_by_id for group_id {group_id}: {e}",
                     exc_info=True,
                 )
-                # For a service method, returning None on DB error might be acceptable,
-                # or raise a custom service-level exception.
                 return None
-            except Exception as e:  # Catch any other unexpected errors
+            except Exception as e:
                 logger.error(
                     f"Unexpected error in get_by_id for group_id {group_id}: {e}",
                     exc_info=True,
@@ -341,7 +341,6 @@ class Service:
         """
         with self.SessionLocal() as session:
             try:
-                # Check if the source statement group exists
                 source_group_exists = (
                     session.query(StatementGroup.id)
                     .filter(StatementGroup.id == group_id)
@@ -352,9 +351,8 @@ class Service:
                         f"Source statement group ID {group_id} not found for "
                         "dependency lookup."
                     )
-                    return None  # Source group does not exist
+                    return None
 
-                # Query for statement groups that `group_id` depends on (citations)
                 cited_target_groups_orm = (
                     session.query(StatementGroup)
                     .join(
@@ -385,7 +383,7 @@ class Service:
                     exc_info=True,
                 )
                 return None
-            except Exception as e:  # Catch any other unexpected errors
+            except Exception as e:
                 logger.error(
                     f"Unexpected error in get_dependencies for "
                     f"group_id {group_id}: {e}",

lean_explore/shared/models/db.py

@@ -21,7 +21,7 @@ from sqlalchemy import (
     MetaData,
     String,
     Text,
-    UniqueConstraint,  # Kept for potential standalone model testing
+    UniqueConstraint,
 )
 from sqlalchemy.orm import DeclarativeBase, Mapped, mapped_column, relationship
 
@@ -66,6 +66,8 @@ class StatementGroup(Base):
             primary declaration.
         informal_description: Optional informal English description, potentially
             LLM-generated.
+        informal_summary: Optional informal English summary, potentially
+            LLM-generated.
         source_file: Relative path to the .lean file containing this block.
         range_start_line: Starting line number of the block in the source file.
         range_start_col: Starting column number of the block.
@@ -95,6 +97,7 @@ class StatementGroup(Base):
     display_statement_text: Mapped[Optional[str]] = mapped_column(Text, nullable=True)
     docstring: Mapped[Optional[str]] = mapped_column(Text, nullable=True)
     informal_description: Mapped[Optional[str]] = mapped_column(Text, nullable=True)
+    informal_summary: Mapped[Optional[str]] = mapped_column(Text, nullable=True)
 
     source_file: Mapped[str] = mapped_column(Text, nullable=False)
     range_start_line: Mapped[int] = mapped_column(Integer, nullable=False)
@@ -169,9 +172,8 @@ class Declaration(Base):
     """Represents a Lean declaration, a node in the dependency graph.
 
     Stores information about Lean declarations (definitions, theorems, axioms, etc.),
-    including source location, Lean code, descriptions, and (potentially)
-    embeddings. Declarations from the same source block can be grouped via
-    `statement_group_id`.
+    including source location, Lean code, and descriptions. Declarations from the
+    same source block can be grouped via `statement_group_id`.
 
     Attributes:
         id: Primary key identifier.
@@ -191,13 +193,6 @@ class Declaration(Base):
         statement_text: Full Lean code text of the originating source block.
         declaration_signature: Extracted Lean signature text of the declaration.
         statement_group_id: Optional foreign key to `statement_groups.id`.
-        type_signature_text: Lean type signature (may be redundant with
-            `declaration_signature`).
-        informal_description: Informal English explanation.
-        lean_embedding: Text representation (e.g., JSON) of the embedding for
-            Lean code.
-        informal_description_embedding: Text representation (e.g., JSON) of the
-            embedding for the informal description.
         pagerank_score: PageRank score within the dependency graph, indexed.
         created_at: Timestamp of record creation.
         updated_at: Timestamp of last record update.
@@ -232,14 +227,6 @@ class Declaration(Base):
         Integer, ForeignKey("statement_groups.id"), nullable=True, index=True
     )
 
-    type_signature_text: Mapped[Optional[str]] = mapped_column(Text, nullable=True)
-    informal_description: Mapped[Optional[str]] = mapped_column(Text, nullable=True)
-
-    lean_embedding: Mapped[Optional[str]] = mapped_column(Text, nullable=True)
-    informal_description_embedding: Mapped[Optional[str]] = mapped_column(
-        Text, nullable=True
-    )
-
     pagerank_score: Mapped[Optional[float]] = mapped_column(
         Float, nullable=True, index=True
     )
@@ -291,7 +278,6 @@ class Dependency(Base):
         source_decl_id: Foreign key to the `Declaration` that depends on another.
         target_decl_id: Foreign key to the `Declaration` that is depended upon.
         dependency_type: String describing the type of dependency (e.g., 'Direct').
-        context: Optional string providing context for the dependency.
         created_at: Timestamp of record creation.
     """
 
@@ -311,7 +297,6 @@ class Dependency(Base):
         index=True,
     )
     dependency_type: Mapped[str] = mapped_column(String(30), nullable=False)
-    context: Mapped[Optional[str]] = mapped_column(String(30), nullable=True)
 
     created_at: Mapped[datetime.datetime] = mapped_column(
         DateTime, default=datetime.datetime.utcnow, nullable=False
@@ -391,7 +376,7 @@ class StatementGroupDependency(Base):
         UniqueConstraint(
             "source_statement_group_id",
             "target_statement_group_id",
-            "dependency_type",  # Consider if type is part of uniqueness
+            "dependency_type",
             name="uq_stmt_group_dependency_link",
         ),
         Index(

{lean_explore-0.1.3.dist-info → lean_explore-0.2.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: lean-explore
-Version: 0.1.3
+Version: 0.2.0
 Summary: A project to explore and rank Lean mathematical declarations.
 Author-email: Justin Asher <justinchadwickasher@gmail.com>
 License:                                  Apache License
@@ -228,8 +228,9 @@ Requires-Dist: sqlalchemy>=2.0
 Requires-Dist: numpy>=1.20
 Requires-Dist: faiss-cpu>=1.7
 Requires-Dist: sentence-transformers>=2.2.0
-Requires-Dist: rapidfuzz>=3.0.0
 Requires-Dist: filelock>=3.0.0
+Requires-Dist: nltk>=3.6
+Requires-Dist: rank-bm25>=0.2.2
 Requires-Dist: httpx>=0.23.0
 Requires-Dist: pydantic>=2.0
 Requires-Dist: typer[all]>=0.9.0
@@ -237,6 +238,7 @@ Requires-Dist: toml>=0.10.0
 Requires-Dist: openai-agents>=0.0.16
 Requires-Dist: mcp>=1.9.0
 Requires-Dist: tqdm>=4.60
+Requires-Dist: requests>=2.25.0
 Dynamic: license-file
 
 # LeanExplore
@@ -261,7 +263,7 @@ If you use LeanExplore in your research or work, please cite it as follows:
 
 **General Citation:**
 
-Justin Asher. (2025). *LeanExplore: A search engine for Lean 4 declarations*. LeanExplore.com. Retrieved from [http://www.leanexplore.com](http://www.leanexplore.com) (GitHub: [https://github.com/justincasher/lean-explore](https://github.com/justincasher/lean-explore)).
+Justin Asher. (2025). *LeanExplore: A search engine for Lean 4 declarations*. LeanExplore.com. (GitHub: [https://github.com/justincasher/lean-explore](https://github.com/justincasher/lean-explore)).
 
 **BibTeX Entry:**
 
@@ -270,7 +272,6 @@ Justin Asher. (2025). *LeanExplore: A search engine for Lean 4 declarations*. Le
   author = {Asher, Justin},
   title = {{LeanExplore: A search engine for Lean 4 declarations}},
   year = {2025},
-  publisher = {LeanExplore.com},
   url = {http://www.leanexplore.com},
   note = {GitHub repository: https://github.com/justincasher/lean-explore}
 }

{lean_explore-0.1.3.dist-info → lean_explore-0.2.0.dist-info}/RECORD

@@ -1,15 +1,15 @@
 lean_explore/__init__.py,sha256=LK4g9wj7jCilTUfQcdQAxDf3F2GjMwzQZJYgnQ8ciGo,38
-lean_explore/defaults.py,sha256=WOg6OpRVusxL8CvHl-kwbQSD66oMa2zf7JCd4f635NY,4498
+lean_explore/defaults.py,sha256=IJw6od-y0grYbwiDJ5ewNZI4u0j0dCCu_AXCDwWLHuA,4459
 lean_explore/api/__init__.py,sha256=LK4g9wj7jCilTUfQcdQAxDf3F2GjMwzQZJYgnQ8ciGo,38
 lean_explore/api/client.py,sha256=AgZG7pUY53Tl1WOhJgUdT0yxa_O1sHsals0pnjRD-Pc,4839
 lean_explore/cli/__init__.py,sha256=LK4g9wj7jCilTUfQcdQAxDf3F2GjMwzQZJYgnQ8ciGo,38
 lean_explore/cli/agent.py,sha256=jf1ebnViAqtKcZAGArqBf9YKHPbsOTpffOXr0Cd2M3Q,29933
 lean_explore/cli/config_utils.py,sha256=_g2aVMM1MyTWJCOVQj_jndI6ZA_Ct09WvpLsP2FFlSM,13643
-lean_explore/cli/data_commands.py,sha256=aIv9-2GdfISZRNtJPv-gCbNCgkuSDulEC4Jt788iSks,19074
+lean_explore/cli/data_commands.py,sha256=mTBqFU7-fF4ZBGzCmNawZA_eHy0jyEMLlBEDEBXpxwY,21462
 lean_explore/cli/main.py,sha256=ZdbXy8x2VQ--JARqJMa9iFnrOhOCLcVgjpWhXkxj80o,24323
 lean_explore/local/__init__.py,sha256=LK4g9wj7jCilTUfQcdQAxDf3F2GjMwzQZJYgnQ8ciGo,38
-lean_explore/local/search.py,sha256=_hnIaJJZLbsgelGHtWS-jb6ojMbZDjCDQZofIfjB8yY,36127
-lean_explore/local/service.py,sha256=bk3-tJM5EVwraPUWhxXBoHjPafd8J-t8qOzr3nZ1w6w,16563
+lean_explore/local/search.py,sha256=ZW8rKJ2riT6RRi6ngo8SylxQ_5jQbsipuv84kqpiwc4,40930
+lean_explore/local/service.py,sha256=7VT_njCiKlMqQItJ-Uy2aDraLttulAAgPhRp2BMWnSk,16166
 lean_explore/mcp/__init__.py,sha256=LK4g9wj7jCilTUfQcdQAxDf3F2GjMwzQZJYgnQ8ciGo,38
 lean_explore/mcp/app.py,sha256=XG6zTAaBRbdV1Ep_Za2JmifEmkYFKBmcGrdizCLlH-s,3808
 lean_explore/mcp/server.py,sha256=pzhLNGfTxelZvQ7ZJrWW0cbNH4MCwhviV24Y-yfQa0c,8666
@@ -17,10 +17,10 @@ lean_explore/mcp/tools.py,sha256=Lri2GNIKNCLZNpNfvgvI600w3-0gaJGdPCFhhd7WVuk,958
 lean_explore/shared/__init__.py,sha256=LK4g9wj7jCilTUfQcdQAxDf3F2GjMwzQZJYgnQ8ciGo,38
 lean_explore/shared/models/__init__.py,sha256=LK4g9wj7jCilTUfQcdQAxDf3F2GjMwzQZJYgnQ8ciGo,38
 lean_explore/shared/models/api.py,sha256=jejNDpgj-cu0KZTqkuOjM0useN4EvhvNB19lFFAOV94,4635
-lean_explore/shared/models/db.py,sha256=tJgct3xGwmvrT0GrsBwlbcz87WvTR1zk_lC9VeNbbUA,16948
-lean_explore-0.1.3.dist-info/licenses/LICENSE,sha256=l4QLw1kIvEOjUktmmKm4dycK1E249Qs2s2AQTYbMXpY,11354
-lean_explore-0.1.3.dist-info/METADATA,sha256=KyF31p7CvkbGFckPT9Eg-rBl8BM2ftoizM17cKS5CFQ,15659
-lean_explore-0.1.3.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-lean_explore-0.1.3.dist-info/entry_points.txt,sha256=JXl2Mo3BRX4jAU-Nxg_CWJR790pB_oi5qnt3Pv5iZnk,58
-lean_explore-0.1.3.dist-info/top_level.txt,sha256=h51BKWrFvB7iym-IlaNAAHX5MZfA8Gmg-aDuXGo0fQ8,13
-lean_explore-0.1.3.dist-info/RECORD,,
+lean_explore/shared/models/db.py,sha256=JYfIBnPrHZO2j7gHAVMlw9WSqVC2NinCG5KuBzdQWyk,16099
+lean_explore-0.2.0.dist-info/licenses/LICENSE,sha256=l4QLw1kIvEOjUktmmKm4dycK1E249Qs2s2AQTYbMXpY,11354
+lean_explore-0.2.0.dist-info/METADATA,sha256=5mCWCIY9zhB7Q2F1Y9mZ9efjKjVPYnQhx2OYP8Gb1sw,15611
+lean_explore-0.2.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+lean_explore-0.2.0.dist-info/entry_points.txt,sha256=JXl2Mo3BRX4jAU-Nxg_CWJR790pB_oi5qnt3Pv5iZnk,58
+lean_explore-0.2.0.dist-info/top_level.txt,sha256=h51BKWrFvB7iym-IlaNAAHX5MZfA8Gmg-aDuXGo0fQ8,13
+lean_explore-0.2.0.dist-info/RECORD,,