feat(pipeline): Anthropic Batch API, source/regulation filter, cost optimization

- Add Anthropic API support to decomposition Pass 0a/0b (prompt caching, content batching)
- Add Anthropic Batch API (50% cost reduction, async 24h processing)
- Add source_filter (ILIKE on source_citation) for regulation-based filtering
- Add category_filter to Pass 0a for selective decomposition
- Add regulation_filter to control_generator for RAG scan phase filtering
  (prefix match on regulation_code — enables CE + Code Review focus)
- New API endpoints: batch-submit-0a, batch-submit-0b, batch-status, batch-process
- 83 new tests (all passing)

Cost reduction: $2,525 → ~$600-700 with all optimizations combined.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

backend-compliance/compliance/services/control_generator.py

@@ -384,6 +384,7 @@ class GeneratorConfig(BaseModel):
     skip_web_search: bool = False
     dry_run: bool = False
     existing_job_id: Optional[str] = None  # If set, reuse this job instead of creating a new one
+    regulation_filter: Optional[List[str]] = None  # Only process chunks matching these regulation_code prefixes
 
 
 @dataclass
@@ -803,6 +804,13 @@ class ControlGeneratorPipeline:
                                 or payload.get("regulation_code", "")
                                 or payload.get("source_id", "")
                                 or payload.get("source_code", ""))
+
+                    # Filter by regulation_code if configured
+                    if config.regulation_filter and reg_code:
+                        code_lower = reg_code.lower()
+                        if not any(code_lower.startswith(f.lower()) for f in config.regulation_filter):
+                            continue
+
                     reg_name = (payload.get("regulation_name_de", "")
                                 or payload.get("regulation_name", "")
                                 or payload.get("source_name", "")