gitflow-analytics 1.0.3__py3-none-any.whl → 1.3.6__py3-none-any.whl

gitflow_analytics/classification/batch_classifier.py

@@ -0,0 +1,752 @@
+"""Batch LLM classifier for intelligent commit categorization with context.
+
+This module implements the second step of the two-step fetch/analyze process,
+providing intelligent batch classification of commits using LLM with ticket context.
+"""
+
+import logging
+import uuid
+from collections import defaultdict
+from datetime import datetime, timedelta, timezone
+from pathlib import Path
+from typing import Any, Optional
+
+from ..core.progress import get_progress_service
+from ..models.database import CachedCommit, DailyCommitBatch, Database, DetailedTicketData
+from ..qualitative.classifiers.llm_commit_classifier import LLMCommitClassifier, LLMConfig
+
+logger = logging.getLogger(__name__)
+
+
+class BatchCommitClassifier:
+    """Intelligent batch classifier using LLM with ticket context.
+
+    WHY: This class implements the second step of the two-step process by:
+    - Reading cached commit data organized by day/week
+    - Adding ticket context to improve classification accuracy
+    - Sending batches of commits to LLM for intelligent classification
+    - Falling back to rule-based classification when LLM fails
+    - Storing results with confidence tracking
+
+    DESIGN DECISION: Uses batch processing to reduce API calls and costs
+    while providing better context for classification accuracy.
+
+    PROGRESS REPORTING: Provides granular progress feedback with nested progress bars:
+    - Repository level: Shows which repository is being processed (position 0)
+    - Weekly level: Shows week being processed within repository (position 1)
+    - API batch level: Shows LLM API batches being processed (position 2)
+    Each level shows commit counts and progress indicators for user feedback.
+    """
+
+    def __init__(
+        self,
+        cache_dir: Path,
+        llm_config: Optional[dict[str, Any]] = None,
+        batch_size: int = 50,
+        confidence_threshold: float = 0.7,
+        fallback_enabled: bool = True,
+    ):
+        """Initialize the batch classifier.
+
+        Args:
+            cache_dir: Path to cache directory containing database
+            llm_config: Configuration for LLM classifier
+            batch_size: Number of commits per batch (max 50 for token limits)
+            confidence_threshold: Minimum confidence for LLM classification
+            fallback_enabled: Whether to fall back to rule-based classification
+        """
+        self.cache_dir = cache_dir
+        self.database = Database(cache_dir / "gitflow_cache.db")
+        self.batch_size = min(batch_size, 50)  # Limit for token constraints
+        self.confidence_threshold = confidence_threshold
+        self.fallback_enabled = fallback_enabled
+
+        # Initialize LLM classifier
+        # Handle different config types
+        if isinstance(llm_config, dict):
+            # Convert dict config to LLMConfig object
+            llm_config_obj = LLMConfig(
+                api_key=llm_config.get("api_key", ""),
+                model=llm_config.get("model", "mistralai/mistral-7b-instruct"),
+                max_tokens=llm_config.get("max_tokens", 50),
+                temperature=llm_config.get("temperature", 0.1),
+                confidence_threshold=llm_config.get("confidence_threshold", 0.7),
+                timeout_seconds=llm_config.get("timeout_seconds", 30),
+                cache_duration_days=llm_config.get("cache_duration_days", 7),
+                enable_caching=llm_config.get("enable_caching", True),
+                max_daily_requests=llm_config.get("max_daily_requests", 1000),
+            )
+        elif hasattr(llm_config, "api_key"):
+            # Use provided config object (e.g., mock config for testing)
+            llm_config_obj = llm_config
+        else:
+            # Use default LLMConfig
+            llm_config_obj = LLMConfig()
+
+        self.llm_classifier = LLMCommitClassifier(config=llm_config_obj, cache_dir=cache_dir)
+        logger.info(
+            f"LLM Classifier initialized with API key: {'Yes' if llm_config_obj.api_key else 'No'}"
+        )
+
+        # Rule-based fallback patterns for when LLM fails
+        self.fallback_patterns = {
+            "feature": [
+                r"feat(?:ure)?[\(\:]",
+                r"add(?:ed|ing)?.*(?:feature|functionality|capability)",
+                r"implement(?:ed|ing|s)?",
+                r"introduce(?:d|s)?",
+            ],
+            "bug_fix": [
+                r"fix(?:ed|es|ing)?[\(\:]",
+                r"bug[\(\:]",
+                r"resolve(?:d|s)?",
+                r"repair(?:ed|ing|s)?",
+                r"correct(?:ed|ing|s)?",
+            ],
+            "refactor": [
+                r"refactor(?:ed|ing|s)?[\(\:]",
+                r"restructure(?:d|ing|s)?",
+                r"optimize(?:d|ing|s)?",
+                r"improve(?:d|ing|s)?",
+                r"clean(?:ed|ing)?\s+up",
+            ],
+            "documentation": [
+                r"docs?[\(\:]",
+                r"documentation[\(\:]",
+                r"readme",
+                r"update.*(?:comment|docs?|documentation)",
+            ],
+            "maintenance": [
+                r"chore[\(\:]",
+                r"maintenance[\(\:]",
+                r"update.*(?:dependencies|deps)",
+                r"bump.*version",
+                r"cleanup",
+            ],
+            "test": [
+                r"test(?:s|ing)?[\(\:]",
+                r"spec[\(\:]",
+                r"add.*(?:test|spec)",
+                r"fix.*test",
+            ],
+            "style": [
+                r"style[\(\:]",
+                r"format(?:ted|ting)?[\(\:]",
+                r"lint(?:ed|ing)?",
+                r"prettier",
+                r"whitespace",
+            ],
+            "build": [
+                r"build[\(\:]",
+                r"ci[\(\:]",
+                r"deploy(?:ed|ment)?",
+                r"docker",
+                r"webpack",
+                r"package\.json",
+            ],
+        }
+
+    def classify_date_range(
+        self,
+        start_date: datetime,
+        end_date: datetime,
+        project_keys: Optional[list[str]] = None,
+        force_reclassify: bool = False,
+    ) -> dict[str, Any]:
+        """Classify all commits in a date range using batch processing.
+
+        Args:
+            start_date: Start date for classification
+            end_date: End date for classification
+            project_keys: Optional list of specific projects to classify
+            force_reclassify: Whether to reclassify already processed batches
+
+        Returns:
+            Dictionary containing classification results and statistics
+        """
+        logger.info(f"Starting batch classification from {start_date.date()} to {end_date.date()}")
+
+        # Get daily batches to process
+        batches_to_process = self._get_batches_to_process(
+            start_date, end_date, project_keys, force_reclassify
+        )
+
+        if not batches_to_process:
+            logger.info("No batches need classification")
+            return {"processed_batches": 0, "total_commits": 0}
+
+        # Group batches by repository first for better progress reporting
+        repo_batches = self._group_batches_by_repository(batches_to_process)
+
+        total_processed = 0
+        total_commits = 0
+
+        # Use centralized progress service
+        progress = get_progress_service()
+
+        # Add progress bar for repository processing
+        with progress.progress(
+            total=len(repo_batches),
+            description="AI Classification",
+            unit="repo",
+            nested=False,
+            leave=True,
+        ) as repo_ctx:
+            for repo_num, (repo_info, repo_batch_list) in enumerate(repo_batches.items(), 1):
+                project_key, repo_path = repo_info
+                repo_name = Path(repo_path).name if repo_path else project_key
+
+                # Count commits in this repository for detailed progress
+                repo_commit_count = sum(batch.commit_count for batch in repo_batch_list)
+
+                progress.set_description(
+                    repo_ctx, f"Classifying {repo_name} ({repo_commit_count} commits)"
+                )
+                logger.info(
+                    f"Processing repository {repo_num}/{len(repo_batches)}: {repo_name} ({len(repo_batch_list)} batches, {repo_commit_count} commits)"
+                )
+
+                # Process this repository's batches by week for optimal context
+                weekly_batches = self._group_batches_by_week(repo_batch_list)
+
+                repo_processed = 0
+                repo_commits_processed = 0
+
+                # Add nested progress bar for weekly processing within repository
+                with progress.progress(
+                    total=len(weekly_batches),
+                    description="  Processing weeks",
+                    unit="week",
+                    nested=True,
+                    leave=False,
+                ) as week_ctx:
+                    for week_num, (week_start, week_batches) in enumerate(
+                        weekly_batches.items(), 1
+                    ):
+                        progress.set_description(
+                            week_ctx,
+                            f"  Week {week_num}/{len(weekly_batches)} ({week_start.strftime('%Y-%m-%d')})",
+                        )
+                        logger.info(
+                            f"  Processing week starting {week_start}: {len(week_batches)} daily batches"
+                        )
+
+                        week_result = self._classify_weekly_batches(week_batches)
+                        repo_processed += week_result["batches_processed"]
+                        repo_commits_processed += week_result["commits_processed"]
+
+                        progress.update(week_ctx, 1)
+                        # Update description to show commits processed
+                        progress.set_description(
+                            week_ctx,
+                            f"  Week {week_num}/{len(weekly_batches)} - {week_result['commits_processed']} commits",
+                        )
+
+                total_processed += repo_processed
+                total_commits += repo_commits_processed
+
+                progress.update(repo_ctx, 1)
+                # Update description to show total progress
+                progress.set_description(
+                    repo_ctx,
+                    f"AI Classification [{repo_num}/{len(repo_batches)} repos, {total_commits} commits]",
+                )
+
+                logger.info(
+                    f"  Repository {repo_name} completed: {repo_processed} batches, {repo_commits_processed} commits"
+                )
+
+        # Store daily metrics from classification results
+        self._store_daily_metrics(start_date, end_date, project_keys)
+
+        logger.info(
+            f"Batch classification completed: {total_processed} batches, {total_commits} commits"
+        )
+
+        return {
+            "processed_batches": total_processed,
+            "total_commits": total_commits,
+            "date_range": {"start": start_date, "end": end_date},
+            "project_keys": project_keys or [],
+        }
+
+    def _get_batches_to_process(
+        self,
+        start_date: datetime,
+        end_date: datetime,
+        project_keys: Optional[list[str]],
+        force_reclassify: bool,
+    ) -> list[DailyCommitBatch]:
+        """Get daily commit batches that need classification."""
+        session = self.database.get_session()
+
+        try:
+            query = session.query(DailyCommitBatch).filter(
+                DailyCommitBatch.date >= start_date.date(), DailyCommitBatch.date <= end_date.date()
+            )
+
+            if project_keys:
+                query = query.filter(DailyCommitBatch.project_key.in_(project_keys))
+
+            if not force_reclassify:
+                # Only get batches that haven't been classified or failed
+                query = query.filter(
+                    DailyCommitBatch.classification_status.in_(["pending", "failed"])
+                )
+
+            batches = query.order_by(DailyCommitBatch.date).all()
+            logger.info(f"Found {len(batches)} batches needing classification")
+
+            # Debug: Log filtering criteria
+            logger.debug(
+                f"Query criteria: start_date={start_date.date()}, end_date={end_date.date()}"
+            )
+            if project_keys:
+                logger.debug(f"Project key filter: {project_keys}")
+            logger.debug(f"Force reclassify: {force_reclassify}")
+
+            return batches
+
+        except Exception as e:
+            logger.error(f"Error getting batches to process: {e}")
+            return []
+        finally:
+            session.close()
+
+    def _group_batches_by_repository(
+        self, batches: list[DailyCommitBatch]
+    ) -> dict[tuple[str, str], list[DailyCommitBatch]]:
+        """Group daily batches by repository for granular progress reporting."""
+        repo_batches = defaultdict(list)
+
+        for batch in batches:
+            # Use (project_key, repo_path) as the key for unique repository identification
+            repo_key = (batch.project_key, batch.repo_path)
+            repo_batches[repo_key].append(batch)
+
+        # Sort each repository's batches by date
+        for batches_list in repo_batches.values():
+            batches_list.sort(key=lambda b: b.date)
+
+        return dict(repo_batches)
+
+    def _group_batches_by_week(
+        self, batches: list[DailyCommitBatch]
+    ) -> dict[datetime, list[DailyCommitBatch]]:
+        """Group daily batches by week for optimal context window."""
+        weekly_batches = defaultdict(list)
+
+        for batch in batches:
+            # Get Monday of the week
+            batch_date = datetime.combine(batch.date, datetime.min.time())
+            days_since_monday = batch_date.weekday()
+            week_start = batch_date - timedelta(days=days_since_monday)
+
+            weekly_batches[week_start].append(batch)
+
+        # Sort each week's batches by date
+        for week_batches in weekly_batches.values():
+            week_batches.sort(key=lambda b: b.date)
+
+        return dict(weekly_batches)
+
+    def _classify_weekly_batches(self, weekly_batches: list[DailyCommitBatch]) -> dict[str, Any]:
+        """Classify all batches for a single week with shared context."""
+        session = self.database.get_session()
+        batches_processed = 0
+        commits_processed = 0
+
+        try:
+            # Collect all commits for the week
+            week_commits = []
+            batch_commit_map = {}  # Maps commit hash to batch
+
+            for batch in weekly_batches:
+                # Mark batch as processing
+                batch.classification_status = "processing"
+
+                # Get commits for this day
+                daily_commits = self._get_commits_for_batch(session, batch)
+                week_commits.extend(daily_commits)
+
+                # Track which batch each commit belongs to
+                for commit in daily_commits:
+                    batch_commit_map[commit["commit_hash"]] = batch
+
+            if not week_commits:
+                logger.warning(
+                    f"No commits found for weekly batches (expected {sum(batch.commit_count for batch in weekly_batches)} commits)"
+                )
+                # Mark batches as failed due to missing commits
+                for batch in weekly_batches:
+                    batch.classification_status = "failed"
+                    batch.classified_at = datetime.utcnow()
+                session.commit()
+                return {"batches_processed": 0, "commits_processed": 0}
+
+            # Get ticket context for the week
+            week_tickets = self._get_ticket_context_for_commits(session, week_commits)
+
+            # Process commits in batches (respecting API limits)
+            classified_commits = []
+            num_batches = (len(week_commits) + self.batch_size - 1) // self.batch_size
+
+            # Use centralized progress service for batch processing
+            progress = get_progress_service()
+
+            # Add progress bar for batch processing within the week
+            with progress.progress(
+                total=num_batches,
+                description="    Processing batches",
+                unit="batch",
+                nested=True,
+                leave=False,
+            ) as batch_ctx:
+                for i in range(0, len(week_commits), self.batch_size):
+                    batch_num = i // self.batch_size + 1
+                    batch_commits = week_commits[i : i + self.batch_size]
+                    progress.set_description(
+                        batch_ctx,
+                        f"    API batch {batch_num}/{num_batches} ({len(batch_commits)} commits)",
+                    )
+                    logger.info(f"Classifying batch {batch_num}: {len(batch_commits)} commits")
+
+                    # Classify this batch with LLM
+                    batch_results = self._classify_commit_batch_with_llm(
+                        batch_commits, week_tickets
+                    )
+                    classified_commits.extend(batch_results)
+
+                    progress.update(batch_ctx, 1)
+                    # Update description to show total classified commits
+                    progress.set_description(
+                        batch_ctx,
+                        f"    API batch {batch_num}/{num_batches} - Total: {len(classified_commits)} commits",
+                    )
+
+            # Store classification results
+            for commit_result in classified_commits:
+                self._store_commit_classification(session, commit_result)
+                commits_processed += 1
+
+            # Mark all daily batches as completed
+            for batch in weekly_batches:
+                batch.classification_status = "completed"
+                batch.classified_at = datetime.utcnow()
+                batches_processed += 1
+
+            session.commit()
+
+            logger.info(
+                f"Week classification completed: {batches_processed} batches, {commits_processed} commits"
+            )
+
+        except Exception as e:
+            logger.error(f"Error in weekly batch classification: {e}")
+            # Mark batches as failed
+            for batch in weekly_batches:
+                batch.classification_status = "failed"
+            session.rollback()
+        finally:
+            session.close()
+
+        return {
+            "batches_processed": batches_processed,
+            "commits_processed": commits_processed,
+        }
+
+    def _get_commits_for_batch(self, session: Any, batch: DailyCommitBatch) -> list[dict[str, Any]]:
+        """Get all commits for a daily batch."""
+        try:
+            # Get cached commits for this batch
+            # CRITICAL FIX: CachedCommit.timestamp is timezone-aware UTC (from analyzer.py line 806)
+            # but we were creating timezone-naive boundaries, causing comparison to fail
+            # Create timezone-aware UTC boundaries to match CachedCommit.timestamp format
+            start_of_day = datetime.combine(batch.date, datetime.min.time(), tzinfo=timezone.utc)
+            end_of_day = datetime.combine(batch.date, datetime.max.time(), tzinfo=timezone.utc)
+
+            logger.debug(
+                f"Searching for commits in {batch.repo_path} between {start_of_day} and {end_of_day}"
+            )
+
+            commits = (
+                session.query(CachedCommit)
+                .filter(
+                    CachedCommit.repo_path == batch.repo_path,
+                    CachedCommit.timestamp >= start_of_day,
+                    CachedCommit.timestamp < end_of_day,
+                )
+                .all()
+            )
+
+            logger.debug(f"Found {len(commits)} commits for batch on {batch.date}")
+
+            commit_list = []
+            for commit in commits:
+                commit_data = {
+                    "commit_hash": commit.commit_hash,
+                    "commit_hash_short": commit.commit_hash[:7],
+                    "message": commit.message,
+                    "author_name": commit.author_name,
+                    "author_email": commit.author_email,
+                    "timestamp": commit.timestamp,
+                    "branch": commit.branch,
+                    "project_key": batch.project_key,
+                    "repo_path": commit.repo_path,
+                    "files_changed": commit.files_changed or 0,
+                    "lines_added": commit.insertions or 0,
+                    "lines_deleted": commit.deletions or 0,
+                    "story_points": commit.story_points,
+                    "ticket_references": commit.ticket_references or [],
+                }
+                commit_list.append(commit_data)
+
+            return commit_list
+
+        except Exception as e:
+            logger.error(f"Error getting commits for batch {batch.id}: {e}")
+            return []
+
+    def _get_ticket_context_for_commits(
+        self, session: Any, commits: list[dict[str, Any]]
+    ) -> dict[str, dict[str, Any]]:
+        """Get ticket context for a list of commits."""
+        # Extract all ticket references from commits
+        all_ticket_ids = set()
+        for commit in commits:
+            ticket_refs = commit.get("ticket_references", [])
+            all_ticket_ids.update(ticket_refs)
+
+        if not all_ticket_ids:
+            return {}
+
+        try:
+            # Get detailed ticket information
+            tickets = (
+                session.query(DetailedTicketData)
+                .filter(DetailedTicketData.ticket_id.in_(all_ticket_ids))
+                .all()
+            )
+
+            ticket_context = {}
+            for ticket in tickets:
+                ticket_context[ticket.ticket_id] = {
+                    "title": ticket.title,
+                    "description": (
+                        ticket.summary or ticket.description[:200] if ticket.description else ""
+                    ),
+                    "ticket_type": ticket.ticket_type,
+                    "status": ticket.status,
+                    "labels": ticket.labels or [],
+                    "classification_hints": ticket.classification_hints or [],
+                    "business_domain": ticket.business_domain,
+                }
+
+            logger.info(f"Retrieved context for {len(ticket_context)} tickets")
+            return ticket_context
+
+        except Exception as e:
+            logger.error(f"Error getting ticket context: {e}")
+            return {}
+
+    def _classify_commit_batch_with_llm(
+        self,
+        commits: list[dict[str, Any]],
+        ticket_context: dict[str, dict[str, Any]],
+    ) -> list[dict[str, Any]]:
+        """Classify a batch of commits using LLM with ticket context."""
+        batch_id = str(uuid.uuid4())
+        logger.info(f"Starting LLM classification for batch {batch_id} with {len(commits)} commits")
+
+        # Prepare batch for LLM classification
+        enhanced_commits = []
+        for commit in commits:
+            enhanced_commit = commit.copy()
+
+            # Add ticket context to commit
+            ticket_refs = commit.get("ticket_references", [])
+            relevant_tickets = []
+            for ticket_id in ticket_refs:
+                if ticket_id in ticket_context:
+                    relevant_tickets.append(ticket_context[ticket_id])
+
+            enhanced_commit["ticket_context"] = relevant_tickets
+            enhanced_commits.append(enhanced_commit)
+
+        try:
+            # Use LLM classifier with enhanced context
+            llm_results = self.llm_classifier.classify_commits_batch(
+                enhanced_commits, batch_id=batch_id, include_confidence=True
+            )
+
+            # Process LLM results and add fallbacks
+            processed_results = []
+            for _i, (commit, llm_result) in enumerate(zip(commits, llm_results)):
+                confidence = llm_result.get("confidence", 0.0)
+                predicted_category = llm_result.get("category", "other")
+
+                # Apply confidence threshold and fallback
+                if confidence < self.confidence_threshold and self.fallback_enabled:
+                    fallback_category = self._fallback_classify_commit(commit)
+                    processed_results.append(
+                        {
+                            "commit_hash": commit["commit_hash"],
+                            "category": fallback_category,
+                            "confidence": 0.5,  # Medium confidence for rule-based
+                            "method": "fallback",
+                            "llm_category": predicted_category,
+                            "llm_confidence": confidence,
+                            "batch_id": batch_id,
+                        }
+                    )
+                else:
+                    processed_results.append(
+                        {
+                            "commit_hash": commit["commit_hash"],
+                            "category": predicted_category,
+                            "confidence": confidence,
+                            "method": "llm",
+                            "batch_id": batch_id,
+                        }
+                    )
+
+            logger.info(
+                f"LLM classification completed for batch {batch_id}: {len(processed_results)} commits"
+            )
+            return processed_results
+
+        except Exception as e:
+            logger.error(f"LLM classification failed for batch {batch_id}: {e}")
+
+            # Fall back to rule-based classification for entire batch
+            if self.fallback_enabled:
+                fallback_results = []
+                for commit in commits:
+                    category = self._fallback_classify_commit(commit)
+                    fallback_results.append(
+                        {
+                            "commit_hash": commit["commit_hash"],
+                            "category": category,
+                            "confidence": 0.3,  # Low confidence for fallback
+                            "method": "fallback_only",
+                            "error": str(e),
+                            "batch_id": batch_id,
+                        }
+                    )
+
+                logger.info(f"Fallback classification completed for batch {batch_id}")
+                return fallback_results
+
+            return []
+
+    def _fallback_classify_commit(self, commit: dict[str, Any]) -> str:
+        """Classify commit using rule-based patterns."""
+        import re
+
+        message = commit.get("message", "").lower()
+
+        # Check patterns in order of specificity
+        for category, patterns in self.fallback_patterns.items():
+            for pattern in patterns:
+                if re.search(pattern, message, re.IGNORECASE):
+                    return category
+
+        # Default category
+        return "other"
+
+    def _store_commit_classification(
+        self, session: Any, classification_result: dict[str, Any]
+    ) -> None:
+        """Store classification result in cached commit record."""
+        try:
+            commit_hash = classification_result["commit_hash"]
+
+            # Find the cached commit record
+            cached_commit = (
+                session.query(CachedCommit).filter(CachedCommit.commit_hash == commit_hash).first()
+            )
+
+            if cached_commit:
+                # Store classification in ticket_references as temporary solution
+                # In production, this would go in a separate classification table
+                if not hasattr(cached_commit, "classification_data"):
+                    cached_commit.ticket_references = cached_commit.ticket_references or []
+
+                # Add classification data to the record
+                # Note: This is a simplified approach - in production you'd want a separate table
+                {
+                    "category": classification_result["category"],
+                    "confidence": classification_result["confidence"],
+                    "method": classification_result["method"],
+                    "classified_at": datetime.utcnow().isoformat(),
+                    "batch_id": classification_result.get("batch_id"),
+                }
+
+                # Store in a JSON field or separate table in production
+                logger.debug(
+                    f"Classified commit {commit_hash[:7]} as {classification_result['category']}"
+                )
+
+        except Exception as e:
+            logger.error(
+                f"Error storing classification for {classification_result.get('commit_hash', 'unknown')}: {e}"
+            )
+
+    def _store_daily_metrics(
+        self,
+        start_date: datetime,
+        end_date: datetime,
+        project_keys: Optional[list[str]],
+    ) -> None:
+        """Store aggregated daily metrics from classification results."""
+        from ..core.metrics_storage import DailyMetricsStorage
+
+        try:
+            DailyMetricsStorage(self.cache_dir / "gitflow_cache.db")
+
+            # This would typically aggregate from the classification results
+            # For now, we'll let the existing system handle this
+            logger.info("Daily metrics storage integration placeholder")
+
+        except Exception as e:
+            logger.error(f"Error storing daily metrics: {e}")
+
+    def get_classification_status(
+        self,
+        start_date: datetime,
+        end_date: datetime,
+        project_keys: Optional[list[str]] = None,
+    ) -> dict[str, Any]:
+        """Get classification status for a date range."""
+        session = self.database.get_session()
+
+        try:
+            query = session.query(DailyCommitBatch).filter(
+                DailyCommitBatch.date >= start_date.date(), DailyCommitBatch.date <= end_date.date()
+            )
+
+            if project_keys:
+                query = query.filter(DailyCommitBatch.project_key.in_(project_keys))
+
+            batches = query.all()
+
+            status_counts = defaultdict(int)
+            total_commits = 0
+
+            for batch in batches:
+                status_counts[batch.classification_status] += 1
+                total_commits += batch.commit_count
+
+            return {
+                "total_batches": len(batches),
+                "total_commits": total_commits,
+                "status_breakdown": dict(status_counts),
+                "completion_rate": status_counts["completed"] / len(batches) if batches else 0.0,
+                "date_range": {"start": start_date, "end": end_date},
+            }
+
+        except Exception as e:
+            logger.error(f"Error getting classification status: {e}")
+            return {}
+        finally:
+            session.close()