Logging¶

DualLens Analytics uses Python's built-in logging module with a centralized configuration function.

Setup¶

Logging is configured once at application startup via duallens_analytics.configure_logging():

from duallens_analytics import configure_logging

configure_logging(level="INFO")

Default format¶

%(asctime)s | %(name)-35s | %(levelname)-8s | %(message)s

Example output:

2025-01-15 10:23:45,123 | duallens_analytics.config          | INFO     | Loading .env from /path/to/.env
2025-01-15 10:23:45,456 | duallens_analytics.vector_store    | INFO     | ChromaDB collection loaded – 142 documents

Per-Module Loggers¶

Every module creates its own logger:

import logging
logger = logging.getLogger(__name__)

This produces a hierarchy under duallens_analytics.*, so you can fine-tune verbosity per module:

logging.getLogger("duallens_analytics.financial").setLevel(logging.WARNING)

What Gets Logged¶

Module	Key Events
`config`	`.env` load, Hydra config resolution, Settings creation
`data_loader`	ZIP extraction, PDF discovery, chunk counts
`vector_store`	Collection build/load, document counts
`financial`	Stock data fetch, metrics fetch, chart generation
`rag`	Question received, context length, answer length
`evaluation`	Judge prompt dispatch, score parsing
`ranking`	Ranking prompt dispatch
`report`	Report generation
`app`	Settings init, vector store init, tab rendering
`cli`	Config resolution output

Disabling Noisy Libraries¶

Third-party libraries (e.g., httpx, chromadb) can be silenced:

logging.getLogger("httpx").setLevel(logging.WARNING)
logging.getLogger("chromadb").setLevel(logging.WARNING)