Files
Anton_wireframe/app/services/querying.py
T

296 lines
12 KiB
Python

import asyncio
import hashlib
import logging
import os
from typing import List, Optional
from db.db import get_db
from db.models import FundTable, InvestorTable, ProjectTable
from langchain_core.prompts import ChatPromptTemplate
from langchain_openai import ChatOpenAI
from schemas.router_schemas import (
CompanyMinimal,
InvestmentResponse,
PaginatedResponse,
SectorMinimal,
)
from sqlalchemy import text
from sqlalchemy.orm import selectinload
from services.compatibility_score import calculate_project_investor_compatibility
logger = logging.getLogger(__name__)
class QueryProcessor:
def __init__(self):
self.llm = ChatOpenAI(
api_key=os.getenv("OPENROUTER_API_KEY"),
base_url="https://openrouter.ai/api/v1",
model="openai/gpt-4o-mini",
temperature=0,
)
# Query cache for performance
self.query_cache = {}
# SQL generation prompt
self.sql_prompt = ChatPromptTemplate.from_messages(
[
(
"system",
"""You are a SQL expert. Generate a SQLite query to find fund IDs based on user requirements.
Database Schema:
- funds: id, fund_name, investor_id, check_size_lower, check_size_upper, geographic_focus
- fund_sectors: fund_id, sector_id
- fund_investment_stages: fund_id, stage_id
- sectors: id, name
- investment_stages: id, name
- investors: id, name, aum
IMPORTANT RULES:
1. ALWAYS return ONLY fund IDs (funds.id) - use SELECT DISTINCT f.id
2. For geography: Be FLEXIBLE - use OR with variations and partial matches
- 'Europe' → WHERE geographic_focus LIKE '%Europe%' OR geographic_focus LIKE '%European%'
- 'America' → WHERE geographic_focus LIKE '%America%' OR geographic_focus LIKE '%US%' OR geographic_focus LIKE '%United States%'
- 'Asia' → WHERE geographic_focus LIKE '%Asia%' OR geographic_focus LIKE '%Asian%'
- If no geography specified, DON'T filter by geography
3. For stages: Use LEFT JOIN and LIKE for flexible matching with synonyms
- 'Seed' → s.name LIKE '%Seed%' OR s.name LIKE '%Pre-Seed%' OR s.name LIKE '%Early%'
- 'Series A' → s.name LIKE '%Series A%' OR s.name LIKE '%A%'
- 'Growth' → s.name LIKE '%Growth%' OR s.name LIKE '%Late%' OR s.name LIKE '%Expansion%'
- If stage not specified, include ALL funds
4. For sectors: Use LEFT JOIN and include related terms with OR
- 'Fintech' → sec.name LIKE '%Fintech%' OR sec.name LIKE '%Finance%' OR sec.name LIKE '%Financial%'
- 'AI' → sec.name LIKE '%AI%' OR sec.name LIKE '%Artificial Intelligence%' OR sec.name LIKE '%Machine Learning%' OR sec.name LIKE '%ML%'
- 'Healthcare' → sec.name LIKE '%Healthcare%' OR sec.name LIKE '%Health%' OR sec.name LIKE '%Medical%' OR sec.name LIKE '%Biotech%'
5. For check size filters (be flexible with ranges):
- "under X" → WHERE (check_size_upper <= X OR check_size_upper IS NULL)
- "over X" → WHERE (check_size_lower >= X OR check_size_lower IS NULL)
- "between X and Y" → WHERE check_size_lower >= X AND check_size_upper <= Y
6. Use LEFT JOIN for stages and sectors so funds without tags still match
7. Use DISTINCT to avoid duplicates from joins
8. Be INCLUSIVE - use OR conditions to cast a wider net
9. If query is very simple (e.g., just "seed stage"), don't add unnecessary filters
10. Return a single, complete SELECT query
Example Queries:
Q: "Seed stage investors in Europe"
A: SELECT DISTINCT f.id FROM funds f
LEFT JOIN fund_investment_stages fis ON f.id = fis.fund_id
LEFT JOIN investment_stages s ON fis.stage_id = s.id
WHERE (s.name LIKE '%Seed%' OR s.name LIKE '%Pre-Seed%' OR s.name LIKE '%Early%' OR s.id IS NULL)
AND (f.geographic_focus LIKE '%Europe%' OR f.geographic_focus LIKE '%European%')
Q: "Fintech investors with check size under 5 million"
A: SELECT DISTINCT f.id FROM funds f
LEFT JOIN fund_sectors fs ON f.id = fs.fund_id
LEFT JOIN sectors sec ON fs.sector_id = sec.id
WHERE (sec.name LIKE '%Fintech%' OR sec.name LIKE '%Finance%' OR sec.name LIKE '%Financial%' OR sec.id IS NULL)
AND (f.check_size_upper <= 5000000 OR f.check_size_upper IS NULL)
Q: "Seed stage investors"
A: SELECT DISTINCT f.id FROM funds f
LEFT JOIN fund_investment_stages fis ON f.id = fis.fund_id
LEFT JOIN investment_stages s ON fis.stage_id = s.id
WHERE s.name LIKE '%Seed%' OR s.name LIKE '%Pre-Seed%' OR s.name LIKE '%Early%'
Q: "Growth stage investors"
A: SELECT DISTINCT f.id FROM funds f
LEFT JOIN fund_investment_stages fis ON f.id = fis.fund_id
LEFT JOIN investment_stages s ON fis.stage_id = s.id
WHERE s.name LIKE '%Growth%' OR s.name LIKE '%Late%' OR s.name LIKE '%Expansion%' OR s.name LIKE '%Series C%' OR s.name LIKE '%Series D%'
Q: "AI investors in America"
A: SELECT DISTINCT f.id FROM funds f
LEFT JOIN fund_sectors fs ON f.id = fs.fund_id
LEFT JOIN sectors sec ON fs.sector_id = sec.id
WHERE (sec.name LIKE '%AI%' OR sec.name LIKE '%Artificial Intelligence%' OR sec.name LIKE '%Machine Learning%' OR sec.name LIKE '%ML%')
AND (f.geographic_focus LIKE '%America%' OR f.geographic_focus LIKE '%US%' OR f.geographic_focus LIKE '%United States%' OR f.geographic_focus LIKE '%USA%')
Q: "Healthcare investors"
A: SELECT DISTINCT f.id FROM funds f
LEFT JOIN fund_sectors fs ON f.id = fs.fund_id
LEFT JOIN sectors sec ON fs.sector_id = sec.id
WHERE sec.name LIKE '%Healthcare%' OR sec.name LIKE '%Health%' OR sec.name LIKE '%Medical%' OR sec.name LIKE '%Biotech%' OR sec.name LIKE '%Pharma%'
IMPORTANT: Use LEFT JOIN so funds without sector/stage tags can still match. Include synonym terms with OR for better recall.
Return ONLY the SQL query, no explanations or markdown.""",
),
("user", "{question}"),
]
)
def _get_cache_key(self, question: str) -> str:
"""Generate cache key from normalized question."""
return hashlib.md5(question.lower().strip().encode()).hexdigest()
async def process_query(
self, question: str, project_id: Optional[int] = None
) -> PaginatedResponse[InvestmentResponse]:
"""Async wrapper for process_query. Runs blocking work in a thread to avoid
blocking the event loop.
"""
return await asyncio.to_thread(self._process_query_sync, question, project_id)
def _process_query_sync(
self, question: str, project_id: Optional[int] = None
) -> PaginatedResponse[InvestmentResponse]:
"""Synchronous implementation of process_query. This is run in a thread by
the async wrapper above.
"""
cache_key = self._get_cache_key(question)
# Check cache first
if cache_key in self.query_cache:
sql_query = self.query_cache[cache_key]
logger.info(f"Using cached SQL: {sql_query}")
else:
# Generate SQL query
messages = self.sql_prompt.format_messages(question=question)
response = self.llm.invoke(messages)
sql_query = response.content.strip()
# Clean up SQL (remove markdown code blocks if present)
sql_query = sql_query.replace("```sql", "").replace("```", "").strip()
# Cache the query
self.query_cache[cache_key] = sql_query
logger.info(f"Generated SQL: {sql_query}")
# Execute query to get fund IDs
db_session = next(get_db())
try:
result = db_session.execute(text(sql_query))
fund_ids = [row[0] for row in result.fetchall()]
logger.info(
f"Found {len(fund_ids)} fund IDs: {fund_ids[:10]}{'...' if len(fund_ids) > 10 else ''}"
)
return self._fetch_funds_by_ids(fund_ids, project_id)
except Exception as e:
logger.error(f"SQL execution error: {e}")
logger.error(f"Failed SQL: {sql_query}")
# Return empty result
return PaginatedResponse(
items=[], total=0, page=1, page_size=10, total_pages=0
)
finally:
db_session.close()
def _fetch_funds_by_ids(
self, fund_ids: List[int], project_id: Optional[int] = None
) -> PaginatedResponse[InvestmentResponse]:
"""Fetch funds with all their relationships from the database using fund IDs.
Constructs response similar to read_investors but starting from funds.
Args:
fund_ids: List of fund IDs to fetch
project_id: Optional project ID for compatibility scoring
"""
if not fund_ids:
return PaginatedResponse(
items=[],
total=0,
page=1,
page_size=len(fund_ids) if fund_ids else 10,
total_pages=0,
)
# Get database session
db_session = next(get_db())
try:
# Load project if project_id provided
project = None
if project_id is not None:
project = (
db_session.query(ProjectTable)
.options(selectinload(ProjectTable.sector))
.filter(ProjectTable.id == project_id)
.first()
)
# Query funds with all necessary relationships loaded
funds = (
db_session.query(FundTable)
.options(
selectinload(FundTable.investor).selectinload(
InvestorTable.portfolio_companies
),
selectinload(FundTable.investor).selectinload(
InvestorTable.team_members
),
selectinload(FundTable.investor).selectinload(
InvestorTable.sectors
),
selectinload(FundTable.investment_stages),
selectinload(FundTable.sectors),
)
.filter(FundTable.id.in_(fund_ids))
.all()
)
# Transform to InvestmentResponse format (one row per fund)
investment_responses = []
for fund in funds:
investor = fund.investor
# Calculate compatibility score if project provided
compatibility_score = 1.0
if project is not None:
compatibility_score = calculate_project_investor_compatibility(
project=project, investor=investor, use_funds=True
)
# Get top 3 portfolio companies (id and name only)
portfolio_companies = [
CompanyMinimal(id=company.id, name=company.name)
for company in investor.portfolio_companies[:3]
]
# Get stage focus as comma-separated string
stage_focus = (
", ".join([stage.name for stage in fund.investment_stages])
if fund.investment_stages
else None
)
# Get top 3 sectors from fund (id and name only) - sorted alphabetically
fund_sectors = [
SectorMinimal(id=sector.id, name=sector.name)
for sector in sorted(fund.sectors[:3] if fund.sectors else [], key=lambda s: s.name)
]
investment_response = InvestmentResponse(
id=investor.id,
name=f"{investor.name} - {fund.fund_name}"
if fund.fund_name
else investor.name,
aum=investor.aum,
check_size_lower=fund.check_size_lower,
check_size_upper=fund.check_size_upper,
geographic_focus=fund.geographic_focus,
stage_focus=stage_focus,
portfolio_companies=portfolio_companies,
sectors=fund_sectors,
compatibility_score=compatibility_score,
)
investment_responses.append(investment_response)
total_count = len(investment_responses)
total_pages = 1 if total_count > 0 else 0
return PaginatedResponse(
items=investment_responses,
total=total_count,
page=1,
page_size=total_count,
total_pages=total_pages,
)
finally:
db_session.close()