feat: Complete cloud-native CIM Document Processor with full BPCP template

🌐 Cloud-Native Architecture:
- Firebase Functions deployment (no Docker)
- Supabase database (replacing local PostgreSQL)
- Google Cloud Storage integration
- Document AI + Agentic RAG processing pipeline
- Claude-3.5-Sonnet LLM integration

✅ Full BPCP CIM Review Template (7 sections):
- Deal Overview
- Business Description
- Market & Industry Analysis
- Financial Summary (with historical financials table)
- Management Team Overview
- Preliminary Investment Thesis
- Key Questions & Next Steps

🔧 Cloud Migration Improvements:
- PostgreSQL → Supabase migration complete
- Local storage → Google Cloud Storage
- Docker deployment → Firebase Functions
- Schema mapping fixes (camelCase/snake_case)
- Enhanced error handling and logging
- Vector database with fallback mechanisms

📄 Complete End-to-End Cloud Workflow:
1. Upload PDF → Document AI extraction
2. Agentic RAG processing → Structured CIM data
3. Store in Supabase → Vector embeddings
4. Auto-generate PDF → Full BPCP template
5. Download complete CIM review

🚀 Generated with [Claude Code](https://claude.ai/code)

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

backend/vector_function.sql

@@ -0,0 +1,32 @@
+-- Enable pgvector extension (if not already enabled)
+CREATE EXTENSION IF NOT EXISTS vector;
+
+-- Create vector similarity search function
+CREATE OR REPLACE FUNCTION match_document_chunks(
+    query_embedding VECTOR(1536),
+    match_threshold FLOAT DEFAULT 0.7,
+    match_count INTEGER DEFAULT 10
+)
+RETURNS TABLE (
+    id UUID,
+    document_id TEXT,
+    content TEXT,
+    metadata JSONB,
+    chunk_index INTEGER,
+    similarity FLOAT
+)
+LANGUAGE SQL STABLE
+AS $$
+    SELECT
+        document_chunks.id,
+        document_chunks.document_id,
+        document_chunks.content,
+        document_chunks.metadata,
+        document_chunks.chunk_index,
+        1 - (document_chunks.embedding <=> query_embedding) AS similarity
+    FROM document_chunks
+    WHERE document_chunks.embedding IS NOT NULL
+    AND 1 - (document_chunks.embedding <=> query_embedding) > match_threshold
+    ORDER BY document_chunks.embedding <=> query_embedding
+    LIMIT match_count;
+$$;