Conapesca Db

MCP server for querying the CONAPESCA historical landings database (2001–2026), covering both the Pacific and Gulf coasts of Mexico.

Exposes tools for extraction and consultation of standardized fishing landings data (avisos de arribo).

Tools available

Data access (raw extraction)

Tool	Description
get_estados(year?)	States with landings, optional year filter
get_species(year?, estado?, tipo_aviso?, top_n?)	Species list with total weight and value
species_count()	Unique scientific names classified by taxonomic level
get_landings(year?, estado?, especie?, tipo_aviso?, oficina?, limit?, group_by?)	Landing records or aggregates; group_by: "year", "estado", "litoral"
get_offices(estado?)	CONAPESCA offices with record counts
get_taxonomy(especie)	Taxonomy + FishBase traits for a species

Reporting (summaries)

Tool	Description
landings_by_year(estado?, tipo_aviso?)	Annual totals: weight, value, species count
landings_by_estado(year?, tipo_aviso?)	Totals by state
landings_by_fleet_type(year?)	MAYORES vs MENORES vs COSECHA

Core

Tool	Description
health_check()	DB connectivity check
schema_snapshot()	Column names, types, row count

Quick start (development)

1. Install

pip install -e ".[dev]"

2. Load dev data (SQLite)

python scripts/load_dev_data.py \
  --csv /path/to/conapesca_landings_2001_2026.csv \
  --db dev/conapesca_dev.sqlite \
  --rows 200000   # optional sample for fast testing

3. Configure

cp .env.example .env
# .env already has USE_SQLITE=true by default

4. Run

python -m mcp_server

Production (MySQL)

Set in .env:

USE_SQLITE=false
CONAPESCA_DB_HOST=...
CONAPESCA_DB_USER=...
CONAPESCA_DB_PASSWORD=...
CONAPESCA_DB_NAME=...

The table expected in MySQL is conapesca_landings_historical with the same schema as the pipeline output.

Resources

• conapesca://data-dictionary — full column descriptions
• conapesca://coverage — temporal and geographic coverage notes