local-deep-research 0.6.0__py3-none-any.whl → 0.6.4__py3-none-any.whl

local_deep_research/web/database/migrations.py

@@ -1,15 +1,18 @@
+import sqlite3
+import uuid
+from pathlib import Path
+
 from loguru import logger
-from sqlalchemy import inspect
+from sqlalchemy import inspect, text, create_engine
 
 from ..services.settings_manager import SettingsManager
 from .models import (
     Base,
-    Journal,
-    Setting,
-    ResearchLog,
-    Research,
-    ResearchHistory,
+    ResearchStrategy,
+    RateLimitAttempt,
+    RateLimitEstimate,
 )
+from ...utilities.db_utils import DB_PATH, get_db_session
 
 
 def import_default_settings_file(db_session):
@@ -33,40 +36,708 @@ def import_default_settings_file(db_session):
 
         # Update the saved version.
         settings_mgr.update_db_version()
+    except Exception:
+        logger.exception("Error importing settings from files")
+
+
+def remove_research_log_table(engine):
+    """
+    Remove the redundant research_log table if it exists
+
+    Args:
+        engine: SQLAlchemy engine
+
+    Returns:
+        bool: True if operation was successful, False otherwise
+    """
+    try:
+        inspector = inspect(engine)
+
+        # Check if table exists
+        if inspector.has_table("research_log"):
+            # For SQLite, we need to use raw SQL for DROP TABLE
+            with engine.connect() as conn:
+                conn.execute("DROP TABLE research_log")
+                conn.commit()
+            logger.info("Successfully removed redundant 'research_log' table")
+            return True
+        else:
+            logger.info("Table 'research_log' does not exist, no action needed")
+            return True
+    except Exception:
+        logger.exception("Error removing research_log table")
+        return False
+
+
+def create_research_strategy_table(engine):
+    """
+    Create the research_strategies table if it doesn't exist
+
+    Args:
+        engine: SQLAlchemy engine
+
+    Returns:
+        bool: True if operation was successful, False otherwise
+    """
+    try:
+        inspector = inspect(engine)
+
+        # Check if table exists
+        if not inspector.has_table("research_strategies"):
+            # Create the table using ORM
+            Base.metadata.create_all(
+                engine, tables=[ResearchStrategy.__table__]
+            )
+            logger.info("Successfully created 'research_strategies' table")
+            return True
+        else:
+            logger.info(
+                "Table 'research_strategies' already exists, no action needed"
+            )
+            return True
+    except Exception:
+        logger.exception("Error creating research_strategies table")
+        return False
+
+
+def create_rate_limiting_tables(engine):
+    """
+    Create rate limiting tables if they don't exist
+
+    Args:
+        engine: SQLAlchemy engine
+
+    Returns:
+        bool: True if operation was successful, False otherwise
+    """
+    try:
+        inspector = inspect(engine)
+
+        tables_to_create = []
+
+        # Check if rate_limit_attempts table exists
+        if not inspector.has_table("rate_limit_attempts"):
+            tables_to_create.append(RateLimitAttempt.__table__)
+            logger.info("Need to create 'rate_limit_attempts' table")
+
+        # Check if rate_limit_estimates table exists
+        if not inspector.has_table("rate_limit_estimates"):
+            tables_to_create.append(RateLimitEstimate.__table__)
+            logger.info("Need to create 'rate_limit_estimates' table")
+
+        if tables_to_create:
+            # Create the tables using ORM
+            Base.metadata.create_all(engine, tables=tables_to_create)
+            logger.info(
+                f"Successfully created {len(tables_to_create)} rate limiting tables"
+            )
+        else:
+            logger.info("Rate limiting tables already exist, no action needed")
+
+        return True
+    except Exception:
+        logger.exception("Error creating rate limiting tables")
+        return False
+
+
+def add_research_id_to_benchmark_results(engine):
+    """
+    Add research_id column to benchmark_results table if it doesn't exist.
+    """
+    try:
+        import sqlite3
+
+        # Get database path from engine
+        db_path = engine.url.database
+
+        logger.info("Checking if benchmark_results needs research_id column...")
+
+        conn = sqlite3.connect(db_path)
+
+        try:
+            cursor = conn.cursor()
+
+            # Check if table exists
+            cursor.execute(
+                "SELECT name FROM sqlite_master WHERE type='table' AND name='benchmark_results'"
+            )
+            if not cursor.fetchone():
+                logger.info("benchmark_results table does not exist, skipping")
+                return True
+
+            # Check if research_id column already exists
+            cursor.execute("PRAGMA table_info(benchmark_results)")
+            columns = cursor.fetchall()
+            has_research_id = any(col[1] == "research_id" for col in columns)
+
+            if has_research_id:
+                logger.info("benchmark_results already has research_id column")
+                return True
+
+            # Add research_id column
+            logger.info("Adding research_id column to benchmark_results table")
+            cursor.execute(
+                "ALTER TABLE benchmark_results ADD COLUMN research_id TEXT"
+            )
+
+            conn.commit()
+            logger.info(
+                "Successfully added research_id column to benchmark_results"
+            )
+            return True
+
+        finally:
+            conn.close()
+
+    except Exception:
+        logger.exception("Error adding research_id column to benchmark_results")
+        return False
+
+
+def add_uuid_id_column_to_research_history(engine):
+    """
+    Adds a new `uuid_id` string column to the `research_history` table if it
+    does not exist already.
+    """
+    try:
+        import sqlite3
+
+        # Get database path from engine
+        db_path = engine.url.database
+
+        logger.info("Checking if research_history needs uuid_id column...")
+
+        conn = sqlite3.connect(db_path)
+
+        try:
+            cursor = conn.cursor()
+
+            # Check if table exists
+            cursor.execute(
+                "SELECT name FROM sqlite_master WHERE type='table' AND name='research_history'"
+            )
+            if not cursor.fetchone():
+                logger.info("research_history table does not exist, skipping")
+                return True
+
+            # Check if uuid_id column already exists
+            cursor.execute("PRAGMA table_info(research_history)")
+            columns = cursor.fetchall()
+            has_uuid_id = any(col[1] == "uuid_id" for col in columns)
+
+            if has_uuid_id:
+                logger.info("research_history already has uuid_id column")
+                return True
+
+            # Add uuid_id column
+            logger.info("Adding uuid_id column to research_history table")
+            cursor.execute(
+                "ALTER TABLE research_history ADD COLUMN uuid_id CHAR(36)"
+            )
+
+            conn.commit()
+            logger.info("Successfully added uuid_id column to research_history")
+            return True
+
+        finally:
+            conn.close()
+
+    except Exception:
+        logger.exception("Error adding uuid_id column to research_history")
+        return False
+
+
+def convert_research_id_to_string_if_needed(engine):
+    """
+    Convert research_id columns from Integer to String in all tables.
+    Preserves existing data by converting integer IDs to string format.
+    Only runs if integer research_id columns are detected.
+    """
+    try:
+        import sqlite3
+
+        # Get database path from engine
+        db_path = engine.url.database
+
+        logger.info(
+            "Checking if research_id columns need conversion to string..."
+        )
+
+        conn = sqlite3.connect(db_path)
+        conn.execute("PRAGMA foreign_keys = OFF")
+
+        try:
+            cursor = conn.cursor()
+
+            # List of tables that might have research_id columns
+            tables_to_check = [
+                "token_usage",
+                "model_usage",
+                "search_calls",
+                "benchmark_results",
+            ]
+
+            tables_needing_conversion = []
+
+            # Check which tables need conversion
+            for table_name in tables_to_check:
+                # Check if table exists
+                cursor.execute(
+                    "SELECT name FROM sqlite_master WHERE type='table' AND name=?",
+                    (table_name,),
+                )
+                if not cursor.fetchone():
+                    continue
+
+                # Check if research_id column exists and is integer type
+                cursor.execute(f"PRAGMA table_info({table_name})")
+                columns = cursor.fetchall()
+
+                for col in columns:
+                    col_name, col_type = col[1], col[2]
+                    if col_name == "research_id" and (
+                        "INTEGER" in col_type.upper()
+                        or "INT" in col_type.upper()
+                    ):
+                        tables_needing_conversion.append(table_name)
+                        break
+
+            if not tables_needing_conversion:
+                logger.info(
+                    "All research_id columns are already string type, no conversion needed"
+                )
+                return True
+
+            logger.info(
+                f"Converting research_id to string in tables: {tables_needing_conversion}"
+            )
+
+            # Convert each table
+            for table_name in tables_needing_conversion:
+                logger.info(f"Converting {table_name} table...")
+
+                # Get the current table schema
+                cursor.execute(
+                    f"SELECT sql FROM sqlite_master WHERE type='table' AND name='{table_name}'"
+                )
+                create_sql = cursor.fetchone()[0]
+
+                # Create new table name
+                new_table_name = f"{table_name}_new"
+
+                # Modify the CREATE TABLE statement to change research_id to TEXT
+                new_create_sql = create_sql.replace(
+                    f"CREATE TABLE {table_name}",
+                    f"CREATE TABLE {new_table_name}",
+                )
+                new_create_sql = new_create_sql.replace(
+                    "research_id INTEGER", "research_id TEXT"
+                )
+                new_create_sql = new_create_sql.replace(
+                    "research_id INT", "research_id TEXT"
+                )
+
+                # Create the new table
+                cursor.execute(new_create_sql)
+
+                # Copy data from old table to new table, converting research_id to string
+                cursor.execute(f"SELECT * FROM {table_name}")
+                old_rows = cursor.fetchall()
+
+                if old_rows:
+                    # Get column names
+                    cursor.execute(f"PRAGMA table_info({table_name})")
+                    columns = cursor.fetchall()
+                    column_names = [col[1] for col in columns]
+                    research_id_index = (
+                        column_names.index("research_id")
+                        if "research_id" in column_names
+                        else -1
+                    )
+
+                    # Prepare insert statement
+                    placeholders = ",".join(["?" for _ in column_names])
+                    insert_sql = f"INSERT INTO {new_table_name} ({','.join(column_names)}) VALUES ({placeholders})"
+
+                    # Convert rows and insert
+                    converted_rows = []
+                    for row in old_rows:
+                        row_list = list(row)
+                        # Convert research_id to string if it's not None
+                        if (
+                            research_id_index >= 0
+                            and row_list[research_id_index] is not None
+                        ):
+                            row_list[research_id_index] = str(
+                                row_list[research_id_index]
+                            )
+                        converted_rows.append(tuple(row_list))
+
+                    cursor.executemany(insert_sql, converted_rows)
+                    logger.info(
+                        f"Converted {len(converted_rows)} rows in {table_name}"
+                    )
+
+                # Drop old table and rename new table
+                cursor.execute(f"DROP TABLE {table_name}")
+                cursor.execute(
+                    f"ALTER TABLE {new_table_name} RENAME TO {table_name}"
+                )
+
+                logger.info(
+                    f"Successfully converted {table_name} research_id to string"
+                )
+
+            # Commit all changes
+            conn.commit()
+            logger.info(
+                "All research_id columns converted to string successfully!"
+            )
+            return True
+
+        finally:
+            conn.execute("PRAGMA foreign_keys = ON")
+            conn.close()
+
+    except Exception:
+        logger.exception("Error converting research_id columns to string")
+        return False
+
+
+def create_provider_models_table(engine):
+    """Create provider_models table for caching available models"""
+    with engine.connect() as conn:
+        result = conn.execute(
+            text(
+                "SELECT name FROM sqlite_master WHERE type='table' AND name='provider_models'"
+            )
+        )
+        if result.fetchone():
+            logger.info(
+                "Table 'provider_models' already exists, no action needed"
+            )
+            return
+
+        logger.info("Creating 'provider_models' table...")
+
+        # Create the table
+        conn.execute(
+            text(
+                """
+            CREATE TABLE provider_models (
+                id INTEGER PRIMARY KEY AUTOINCREMENT,
+                provider VARCHAR(50) NOT NULL,
+                model_key VARCHAR(255) NOT NULL,
+                model_label VARCHAR(255) NOT NULL,
+                model_metadata JSON,
+                last_updated DATETIME DEFAULT CURRENT_TIMESTAMP NOT NULL,
+                UNIQUE(provider, model_key)
+            )
+            """
+            )
+        )
+
+        # Create index on provider
+        conn.execute(
+            text(
+                "CREATE INDEX ix_provider_models_provider ON provider_models (provider)"
+            )
+        )
+
+        conn.commit()
+        logger.info("Table 'provider_models' created successfully")
+
+
+def migrate_to_uuid():
+    """
+    Migrate all research_id fields from integers to UUIDs.
+
+    Strategy:
+    1. Add new UUID columns alongside existing integer columns
+    2. Generate UUIDs for existing data (or keep as string versions of integers)
+    3. Update foreign key relationships
+    4. Drop old integer columns and rename UUID columns
+    """
+    db_path = Path(DB_PATH)
+    if not db_path.exists():
+        logger.info("Database doesn't exist yet, migration not needed")
+        return
+
+    logger.info(f"Starting UUID migration on {db_path}")
+
+    conn = sqlite3.connect(db_path)
+    conn.execute(
+        "PRAGMA foreign_keys = OFF"
+    )  # Disable FK constraints during migration
+
+    try:
+        cursor = conn.cursor()
+
+        # 1. Migrate research_history table (main research IDs)
+        logger.info("Migrating research_history table...")
+
+        # Check if the table exists and has the old structure
+        cursor.execute("PRAGMA table_info(research_history)")
+        columns = cursor.fetchall()
+        has_uuid_id = any(col[1] == "uuid_id" for col in columns)
+
+        if not has_uuid_id:
+            # Add UUID column
+            cursor.execute(
+                "ALTER TABLE research_history ADD COLUMN uuid_id TEXT"
+            )
+
+            # Generate UUIDs for existing records (convert integer ID to UUID format)
+            cursor.execute("SELECT id FROM research_history")
+            existing_ids = cursor.fetchall()
+
+            for (old_id,) in existing_ids:
+                # Generate a deterministic UUID based on the old ID
+                new_uuid = str(
+                    uuid.uuid5(uuid.NAMESPACE_OID, f"research_{old_id}")
+                )
+                cursor.execute(
+                    "UPDATE research_history SET uuid_id = ? WHERE id = ?",
+                    (new_uuid, old_id),
+                )
+
+            logger.info(
+                f"Generated UUIDs for {len(existing_ids)} research records"
+            )
+
+        # 2. Migrate metrics tables
+        logger.info("Migrating metrics tables...")
+
+        # Token usage table
+        cursor.execute("PRAGMA table_info(token_usage)")
+        columns = cursor.fetchall()
+        has_uuid_research_id = any(
+            col[1] == "uuid_research_id" for col in columns
+        )
+
+        if not has_uuid_research_id:
+            cursor.execute(
+                "ALTER TABLE token_usage ADD COLUMN uuid_research_id TEXT"
+            )
+
+            # Convert existing research_ids to UUIDs (deterministic conversion)
+            cursor.execute(
+                "SELECT DISTINCT research_id FROM token_usage WHERE research_id IS NOT NULL"
+            )
+            research_ids = cursor.fetchall()
+
+            for (research_id,) in research_ids:
+                if research_id:
+                    new_uuid = str(
+                        uuid.uuid5(
+                            uuid.NAMESPACE_OID, f"research_{research_id}"
+                        )
+                    )
+                    cursor.execute(
+                        "UPDATE token_usage SET uuid_research_id = ? WHERE research_id = ?",
+                        (new_uuid, research_id),
+                    )
+
+            logger.info(
+                f"Migrated {len(research_ids)} research IDs in token_usage"
+            )
+
+        # Model usage table
+        cursor.execute("PRAGMA table_info(model_usage)")
+        columns = cursor.fetchall()
+        has_uuid_research_id = any(
+            col[1] == "uuid_research_id" for col in columns
+        )
+
+        if not has_uuid_research_id:
+            cursor.execute(
+                "ALTER TABLE model_usage ADD COLUMN uuid_research_id TEXT"
+            )
+
+            cursor.execute(
+                "SELECT DISTINCT research_id FROM model_usage WHERE research_id IS NOT NULL"
+            )
+            research_ids = cursor.fetchall()
+
+            for (research_id,) in research_ids:
+                if research_id:
+                    new_uuid = str(
+                        uuid.uuid5(
+                            uuid.NAMESPACE_OID, f"research_{research_id}"
+                        )
+                    )
+                    cursor.execute(
+                        "UPDATE model_usage SET uuid_research_id = ? WHERE research_id = ?",
+                        (new_uuid, research_id),
+                    )
+
+            logger.info(
+                f"Migrated {len(research_ids)} research IDs in model_usage"
+            )
+
+        # Search calls table
+        cursor.execute("PRAGMA table_info(search_calls)")
+        columns = cursor.fetchall()
+        has_uuid_research_id = any(
+            col[1] == "uuid_research_id" for col in columns
+        )
+
+        if not has_uuid_research_id:
+            cursor.execute(
+                "ALTER TABLE search_calls ADD COLUMN uuid_research_id TEXT"
+            )
+
+            cursor.execute(
+                "SELECT DISTINCT research_id FROM search_calls WHERE research_id IS NOT NULL"
+            )
+            research_ids = cursor.fetchall()
+
+            for (research_id,) in research_ids:
+                if research_id:
+                    new_uuid = str(
+                        uuid.uuid5(
+                            uuid.NAMESPACE_OID, f"research_{research_id}"
+                        )
+                    )
+                    cursor.execute(
+                        "UPDATE search_calls SET uuid_research_id = ? WHERE research_id = ?",
+                        (new_uuid, research_id),
+                    )
+
+            logger.info(
+                f"Migrated {len(research_ids)} research IDs in search_calls"
+            )
+
+        # 3. Migrate benchmark tables
+        logger.info("Migrating benchmark tables...")
+
+        # Check if benchmark_results table exists
+        cursor.execute(
+            "SELECT name FROM sqlite_master WHERE type='table' AND name='benchmark_results'"
+        )
+        if cursor.fetchone():
+            cursor.execute("PRAGMA table_info(benchmark_results)")
+            columns = cursor.fetchall()
+            has_uuid_research_id = any(
+                col[1] == "uuid_research_id" for col in columns
+            )
+
+            if not has_uuid_research_id:
+                cursor.execute(
+                    "ALTER TABLE benchmark_results ADD COLUMN uuid_research_id TEXT"
+                )
+
+                cursor.execute(
+                    "SELECT DISTINCT research_id FROM benchmark_results WHERE research_id IS NOT NULL"
+                )
+                research_ids = cursor.fetchall()
+
+                for (research_id,) in research_ids:
+                    if research_id:
+                        new_uuid = str(
+                            uuid.uuid5(
+                                uuid.NAMESPACE_OID, f"research_{research_id}"
+                            )
+                        )
+                        cursor.execute(
+                            "UPDATE benchmark_results SET uuid_research_id = ? WHERE research_id = ?",
+                            (new_uuid, research_id),
+                        )
+
+                logger.info(
+                    f"Migrated {len(research_ids)} research IDs in benchmark_results"
+                )
+
+        # Commit all changes
+        conn.commit()
+        logger.info("UUID migration completed successfully!")
+
+        # Note: We're keeping both old and new columns for now
+        # The application will use the new UUID columns
+        # Old columns can be dropped in a future migration once everything is stable
+
     except Exception as e:
-        logger.error("Error importing settings from files: %s", e)
+        logger.error(f"Error during UUID migration: {e}")
+        conn.rollback()
+        raise
+    finally:
+        conn.execute("PRAGMA foreign_keys = ON")  # Re-enable FK constraints
+        conn.close()
+
+
+def cleanup_old_columns():
+    """
+    Cleanup migration - drops old integer columns after UUID migration is stable.
+    Run this only after confirming the UUID migration is working correctly.
+    """
+    logger.warning(
+        "This will permanently remove old integer research_id columns!"
+    )
+    logger.warning("Make sure to backup your database before running this!")
+
+    conn = sqlite3.connect(DB_PATH)
+    conn.execute("PRAGMA foreign_keys = OFF")
+
+    try:
+        # For SQLite, we need to recreate tables to drop columns
+        # This is complex, so we'll leave old columns for now
+        # They can be cleaned up manually if needed
+
+        logger.info("Cleanup deferred - old columns remain for safety")
 
+    finally:
+        conn.execute("PRAGMA foreign_keys = ON")
+        conn.close()
 
-def run_migrations(engine, db_session=None):
+
+def ensure_database_initialized(engine=None, db_session=None):
     """
-    Run any necessary database migrations
+    Initialize the database and run any necessary database migrations
 
     Args:
         engine: SQLAlchemy engine
         db_session: Optional SQLAlchemy session
     """
+    logger.info("Initializing database at {}.", DB_PATH)
+
+    if engine is None:
+        # Create SQLAlchemy engine
+        engine = create_engine(f"sqlite:///{DB_PATH}")
+    if db_session is None:
+        db_session = get_db_session()
+
     # Create all tables if they don't exist
-    inspector = inspect(engine)
-    if not inspector.has_table("settings"):
-        logger.info("Creating settings table")
-        Base.metadata.create_all(engine, tables=[Setting.__table__])
+    Base.metadata.create_all(engine)
+
+    # Import existing settings from files
+    import_default_settings_file(db_session)
 
-    if not inspector.has_table(Journal.__tablename__):
-        logger.info("Creating journals table.")
-        Base.metadata.create_all(engine, tables=[Journal.__table__])
+    try:
+        # 1. Remove the redundant research_log table
+        remove_research_log_table(engine)
 
-    if not inspector.has_table(ResearchLog.__tablename__):
-        logger.info("Creating research logs table.")
-        Base.metadata.create_all(engine, tables=[ResearchLog.__table__])
+        # 2. Create research_strategies table
+        create_research_strategy_table(engine)
 
-    if not inspector.has_table(Research.__tablename__):
-        logger.info("Creating research table.")
-        Base.metadata.create_all(engine, tables=[Research.__table__])
+        # 4. Create rate limiting tables
+        create_rate_limiting_tables(engine)
 
-    if not inspector.has_table(ResearchHistory.__tablename__):
-        logger.info("Creating research table.")
-        Base.metadata.create_all(engine, tables=[ResearchHistory.__table__])
+        # 5. Add research_id column to benchmark_results if missing
+        add_research_id_to_benchmark_results(engine)
 
-    # Import existing settings from files
-    if db_session:
-        import_default_settings_file(db_session)
+        # 6. Convert research_id columns from integer to string
+        convert_research_id_to_string_if_needed(engine)
+
+        # 7. Add uuid_id column to research_history if missing
+        add_uuid_id_column_to_research_history(engine)
+
+        # 8. Create provider_models table for caching
+        create_provider_models_table(engine)
+
+        logger.info("Schema upgrades completed successfully")
+        return True
+    except Exception:
+        logger.exception("Error during schema upgrades")
+        return False
+    finally:
+        db_session.close()