Python client for accessing and analyzing macroeconomic data from the Banco Central de Reserva del Perú (BCRP).

The BCRP provides a public API with hundreds of economic time series, including:

• Inflation (IPC)
• GDP and economic activity
• Interest rates
• Monetary aggregates
• External sector indicators

Official API: 👉 https://estadisticas.bcrp.gob.pe/estadisticas/series/ayuda/api

Explore data: 👉 https://estadisticas.bcrp.gob.pe/estadisticas/series/

bcrpy is a lightweight data access layer for BCRP macroeconomic data.

It provides:

• Simple Python interface (get, large_get)
• Parallel data fetching
• Built-in caching (disk + SQLite)
• Metadata exploration tools
• Clean pandas output

git clone https://github.com/andrewrgarcia/bcrpy
cd bcrpy
uv sync

from bcrpy import get

df = get(
    codes=["PN01288PM"],  # inflation
    start="2020-01",
    end="2023-01"
)

print(df.head())

from bcrpy import large_get

df = large_get(
    codes=["PN01288PM", "PN01289PM", "PN00015MM"],
    start="2010-01",
    end="2023-01",
    chunk_size=2,
    workers=4
)

from bcrpy import Marco

m = Marco()
m.get_metadata()

tools = m.tools()

tools.query("PN01288PM")
tools.wordsearch("inflacion")

Data is cached automatically:

.bcrpy_cache/
├── cache_<hash>.bcrfile
├── cache_<hash>.meta
├── large_cache_<hash>.bcrfile

Features:

• Automatic reuse
• Hash-based isolation (no collisions)
• Optional SQLite backend

get(..., forget=True)     # ignore cache
large_get(..., workers=2) # control parallelism

Run all tests:

uv run pytest -v

bcrpy/
├── _fetcher.py     # core data retrieval
├── _http.py        # HTTP + retry logic
├── _metadata.py    # metadata handling
├── tools.py        # search / query tools
├── utils.py        # IO + helpers

• Minimal dependencies
• Explicit behavior (no hidden magic)
• Reproducibility via caching
• Separation of concerns (data vs tools)

• No API key required (public BCRP API)
• Network-dependent (uses official endpoint)
• Designed for research and modeling workflows

MIT