Internacia Python SDK

A Python SDK for accessing internacia-db data with support for countries, international blocks, and fuzzy search across multiple languages.

Features

• 🌍 Countries Data: Access comprehensive country information including codes, names, currencies, languages, and more
• 🏛️ International Blocks: Query international organizations, alliances, and unions
• 🔍 Fuzzy Search: Multi-language search supporting partial matches across names, codes, acronyms, and names in other languages
• 🚀 Thread-Safe: Built on DuckDB with thread-safe connection handling
• 🌐 Multi-Language: Search works across all languages supported in the dataset
• ⚡ Fast: Leverages DuckDB for high-performance queries
• ✅ Well-Tested: Comprehensive test suite with 100% coverage of all public APIs

Compatibility

internacia (SDK)	internacia-db data	duckdb	Python
0.4.x	3.x (contract 3.x; tested against 3.1.0)	>=1.5.4,<3	>=3.11
0.3.x	3.x (tested against 3.0.0)	>=1.5.4,<3	>=3.11
0.2.0	1.x–2.x; broken on 3.x	>=0.9.0	>=3.8 (declared)

Installation

pip install internacia

For download progress bars, install with the progress extra:

pip install internacia[progress]

For the MCP server (internacia-mcp), install the mcp extra:

pip install internacia[mcp]

Or install from source:

git clone https://github.com/datenoio/internacia-python.git
cd internacia-python
pip install -e .

Getting the database

The SDK reads an internacia.duckdb file from an internacia-db release. Download it once (similar to NLTK's download mechanism):

from internacia import InternaciaClient, download_database

download_database()                    # newest release with a supported contract
download_database(version="3.0.0")     # a specific release ("v3.0.0" also works)
download_database(force=True)          # re-download even if cached
download_database(show_progress=True)  # progress bar (pip install internacia[progress])

client = InternaciaClient()            # uses the newest cached release

• Files come from direct release download URLs, so downloading does not hit the GitHub API rate limit. The API is only used by list_releases(), or when the latest release has a newer contract major; set GITHUB_TOKEN to authenticate those calls.
• Each download is checked against the release's SHA256SUMS. Releases published before SHA256SUMS existed (v3.0.0) are downloaded with a warning.
• Releases are cached in ~/.internacia/<tag>/internacia.duckdb (override with INTERNACIA_CACHE_DIR). The client picks the newest cached release by semantic version.

Pin a release with InternaciaClient(data_version="3.0.0") or INTERNACIA_DATA_VERSION; the same variable is the default version for download_database(). To use any other file (for example a local python scripts/builder.py build in internacia-db), pass db_path= or set INTERNACIA_DB_PATH.

from internacia import check_for_updates, cached_versions, get_latest_version

update = check_for_updates()   # compares the newest cached release with the newest release
if update["has_update"]:
    download_database(update["latest_version"])

print(get_latest_version(), cached_versions())

Version and contract

meta = client.meta()
meta["version"]            # "3.0.0"
meta["contract_version"]   # "3.0"
meta["datasets"]["countries"]  # {"row_count": 256, "schema_hash": "..."}

internacia-db versions its schema as a data contract (contract_version = MAJOR.MINOR). This SDK release supports contract 3.x and was tested against 3.0:

• A different MAJOR raises ContractError when the client is created.
• A newer MINOR (new tables, columns or enum values) is accepted with a logged warning.
• InternaciaClient(check_contract=False) skips the check.
• Enum-like fields (status, scope_category, geographic_scope, …) are open strings: new values can appear in any MINOR release.

Memberships, block types and retired IDs

with InternaciaClient() as client:   # closes the connection on exit
    # Current membership edges (no former members, no left date, block not dissolved)
    nato = client.memberships.get_by_intblock("NATO")
    full = [m["country_code"] for m in nato if m["status"] in ("member", "founding_member")]
    history = client.memberships.get_by_country("US", current_only=False)
    client.memberships.is_member("US", "NAFTA")             # False (left in 2020)

    client.intblocks.get_by_member("US", current_only=True)  # blocks, current only
    client.blocktypes.get_all()                              # the blocktype taxonomy

    client.intblocks.resolve_id("USAN")    # {"id": "UNASUR", "status": "merged", ...}
    client.countries.resolve_code("KV")    # {"id": "XK", "status": "renamed", ...}

resolve_id returns status current, an alias reason (renamed, merged, retired) with the replacement id (None when retired), or unknown. disambiguated is set when the id is current but used to name another entity. Alias resolution needs an internacia-db release that ships the alias tables inside internacia.duckdb (after 3.0.0); with 3.0.0 only current and unknown are reported.

Derived catalogs

Currencies, languages, time zones and regions come from the canonical views of the data contract (or the same SQL on databases built before the views existed). Countries with no currencies, languages or time zones simply have no rows.

client.countries.get_currencies()               # [{"code", "name", "symbol", "country_count"}]
client.countries.get_currency("EUR")
client.countries.get_country_currencies("CH")   # [{"code": "CHF", ...}]
client.countries.get_languages(); client.countries.get_country_languages("BE")
client.countries.get_timezones(); client.countries.get_country_timezones("RU")
client.countries.get_by_timezone("Europe/Paris")
client.countries.get_regions(); client.countries.get_region("ECS")

LLM tools

internacia.tools turns the SDK into function-calling tools. Every public query method is a tool (countries_get_by_code, memberships_get_by_intblock, …) whose schema is generated from its signature and docstring, plus two composite tools, country_lookup and intblock_lookup. Passing a client fills enum lists (regions, statuses, block types, scope categories) from the connected data version.

from internacia import InternaciaClient
from internacia.tools import call_tool, openai_tools, tool_schemas

client = InternaciaClient()
tools = openai_tools(client)        # OpenAI "tools" format; tool_schemas() for plain JSON Schema

# After the model returns a tool call:
result = call_tool(client, "intblock_lookup", '{"member_country": "FR", "scope_category": "igo"}')

call_tool accepts the arguments as a dict or as the JSON string LLM APIs return, and returns JSON-serializable data.

MCP server

The same tools are available to MCP clients (Claude Desktop, Cursor, …) through a stdio server:

pip install internacia[mcp]
internacia-mcp

{"mcpServers": {"internacia": {"command": "internacia-mcp"}}}

The server finds the database like the client does (INTERNACIA_DB_PATH, INTERNACIA_DATA_VERSION or the download cache); set them in the client's env block. internacia.mcp_server.build_server(client) returns the server for embedding.

Quick Start

from internacia import InternaciaClient

# Initialize the client
client = InternaciaClient()

# Get a country by code
country = client.countries.get_by_code("US")
print(country["name"])  # "United States"

# Search for countries or blocks
results = client.search.fuzzy("United States")
for result in results:
    print(f"{result['type']}: {result['name']}")

# Get an international block
block = client.intblocks.get_by_id("EU")
print(block["name"])  # "European Union"

Usage

Countries

Get Country by Code

# By ISO 3166-1 alpha-2 code
country = client.countries.get_by_code("US")

# By ISO 3166-1 alpha-3 code
country = client.countries.get_by_iso3("USA")

# By numeric code
country = client.countries.get_by_numeric_code("840")

Query Countries

# Get all countries
all_countries = client.countries.get_all()

# Get UN members
un_members = client.countries.get_un_members()

# Get independent countries
independent = client.countries.get_independent()

# Get countries by region
countries = client.countries.get_by_region("NAC")  # North America

# Get countries by income level
countries = client.countries.get_by_income_level("OEC")  # High income: OECD

# Get countries by continent
countries = client.countries.get_by_continent("Europe")

# Get countries by currency
countries = client.countries.get_by_currency("EUR")

# Get countries by language
countries = client.countries.get_by_language("eng")

# Get total count
count = client.countries.count()

International Blocks

Get Block by ID

block = client.intblocks.get_by_id("EU")

Query Blocks

# Get all blocks
all_blocks = client.intblocks.get_all()

# Get blocks by type
blocks = client.intblocks.get_by_blocktype("economic")

# Get blocks by status
blocks = client.intblocks.get_by_status("formal")

# Get blocks by geographic scope
blocks = client.intblocks.get_by_geographic_scope("regional")

# Get blocks containing a specific country
blocks = client.intblocks.get_by_member("US")

# Get blocks by acronym
blocks = client.intblocks.get_by_acronym("EU")

# Get blocks by tag
blocks = client.intblocks.get_by_tag("trade")

# Get blocks by topic
blocks = client.intblocks.get_by_topic("economy")

# Get blocks founded in a specific year
blocks = client.intblocks.get_by_founded_year(1993)

# Get total count
count = client.intblocks.count()

Search

Fuzzy Search

The fuzzy search function searches across countries and international blocks, supporting:

• Country names (in any language via native_names)
• Country codes (ISO 3166-1 alpha-2, alpha-3, numeric)
• International block names
• Block names in other languages (other_names)
• Block acronyms (in any language)
• Block IDs
• Block tags

# Search across both countries and blocks
results = client.search.fuzzy("United States")

# Search only countries
countries = client.search.search_countries("United")

# Search only blocks
blocks = client.search.search_intblocks("EU")

# Search with custom limit
results = client.search.fuzzy("Europe", limit=5)

# Control search scope
results = client.search.fuzzy("EU", search_countries=False, search_intblocks=True)

# Multi-language search
results = client.search.fuzzy("Европа")  # Russian
results = client.search.fuzzy("欧盟")    # Chinese
results = client.search.fuzzy("UE")      # French acronym

The fuzzy() method supports the following parameters:

• query: Search query string (required)
• limit: Maximum number of results to return (default: 10)
• search_countries: Whether to search countries (default: True)
• search_intblocks: Whether to search international blocks (default: True)

Results are ranked: exact codes and block ids first, then names (or codes and ids) that start with the query, then other matches; ties go to shorter names. Each table is searched with one query. % and _ in the query match literally.

Custom Database Path

If your database is in a different location:

from pathlib import Path

client = InternaciaClient(
    db_path=Path("/path/to/internacia.duckdb")
)

Configuration

The SDK supports multiple ways to configure database paths and logging:

Environment Variables

# Set custom database path
export INTERNACIA_DB_PATH="/path/to/internacia.duckdb"

# Set log level (DEBUG, INFO, WARNING, ERROR, CRITICAL)
export INTERNACIA_LOG_LEVEL="DEBUG"

# Set custom cache directory for downloaded databases
export INTERNACIA_CACHE_DIR="/path/to/cache"

# Pin a cached release (also the default for download_database())
export INTERNACIA_DATA_VERSION="3.0.0"

# Authenticate GitHub API calls (list_releases)
export GITHUB_TOKEN="..."

The SDK will automatically use these environment variables if set. Priority order for database path resolution:

1. Explicit db_path parameter
2. INTERNACIA_DB_PATH environment variable
3. The cached release pinned by data_version / INTERNACIA_DATA_VERSION
4. The newest cached release with a supported contract major

Logging

The SDK logs through the standard logging module under the internacia logger and never attaches handlers of its own (the package logger has a NullHandler). Configure output in your application:

import logging

logging.basicConfig(level=logging.INFO)
logging.getLogger("internacia").setLevel(logging.DEBUG)  # queries and row counts

INTERNACIA_LOG_LEVEL (DEBUG, INFO, WARNING, ERROR, CRITICAL) sets the level of the internacia logger when a client is created.

Type Hints

The SDK provides comprehensive type hints using TypedDict models for better IDE support and type checking:

from internacia import InternaciaClient, Country, Intblock, SearchResult

client = InternaciaClient()

# All methods return properly typed dictionaries
country: Country = client.countries.get_by_code("US")
block: Intblock = client.intblocks.get_by_id("EU")
results: list[SearchResult] = client.search.fuzzy("United States")

Available types:

• Country: Country data structure
• Intblock: International block data structure
• SearchResult: Search result with type indicator
• Supporting types: CapitalCity, Region, IncomeLevel, Language, Currency, NativeName, OtherName, Acronym, Topic, Member

Error Handling

The SDK provides a custom exception hierarchy for better error handling:

from internacia import InternaciaClient
from internacia.exceptions import (
    InternaciaError,
    DatabaseError,
    NotFoundError,
    ValidationError,
    DownloadError,
    VersionError,
)

client = InternaciaClient()

try:
    country = client.countries.get_by_code("XX")  # Invalid code
except ValidationError as e:
    print(f"Invalid input: {e}")
except NotFoundError as e:
    print(f"Not found: {e}")
except DatabaseError as e:
    print(f"Database error: {e}")

Exception types:

• InternaciaError: Base exception for all SDK errors
• DatabaseError: Database connection or query errors
• NotFoundError: Resource not found (e.g., country code doesn't exist)
• ValidationError: Input validation errors (e.g., empty query string)
• DownloadError: Database download or checksum verification failures
• VersionError: Version-related errors
• ContractError (a VersionError): the database's data contract major is not supported

Data Structure

Country Fields

• code: ISO 3166-1 alpha-2 code (e.g., "US")
• name: Common name
• iso3code: ISO 3166-1 alpha-3 code
• numeric_code: ISO 3166-1 numeric code
• official_name: Official full name
• capital_city: {name, lng, lat}
• region: World Bank region {id, value}
• incomeLevel: World Bank income level {id, value}
• languages: List of {code, name, official}
• currencies: List of {code, name, symbol}
• un_member: Boolean
• independent: Boolean
• continents: List of continent names
• borders: List of bordering country codes
• native_names: Map of language code -> {official, common}
• And more...

International Block Fields

• id: Unique identifier
• name: Name of the block
• blocktype: List of types (e.g., ["economic", "political"])
• status: Status ("formal", "informal", "de-facto")
• other_names: List of {id, name}, where id is the language code
• acronyms: List of {lang, value}
• includes: List of member countries {id, name, type, status, ...}
• founded: Foundation year/date
• geographic_scope: Scope ("global", "regional", "sub-regional")
• scope_category: Scope category (an open string; new values can appear in any data release)
• tags: List of classification tags
• topics: List of {key, name}
• And more...

Thread Safety

The SDK is designed to be thread-safe. Each query uses its own read-only DuckDB connection, which is automatically closed after use. This allows safe concurrent access from multiple threads.

import threading
from internacia import InternaciaClient

client = InternaciaClient()

def query_country(code):
    country = client.countries.get_by_code(code)
    print(f"{code}: {country['name']}")

# Safe to use from multiple threads
threads = []
for code in ["US", "FR", "DE", "GB"]:
    t = threading.Thread(target=query_country, args=(code,))
    threads.append(t)
    t.start()

for t in threads:
    t.join()

Examples

See the examples/ directory for more detailed usage examples:

• basic_usage.py: Basic operations
• search_examples.py: Search functionality
• advanced_queries.py: Complex queries
• multi_language.py: Multi-language search examples

Development

Setup

git clone https://github.com/datenoio/internacia-python.git
cd internacia-python
pip install -e ".[dev]"

Running Tests

The suite has three parts:

• Unit tests (tests/test_*.py) mock the database and cover validation, error paths and the downloader.
• Integration tests (tests/integration/) run every public method and every examples/*.py script against a real internacia-db release DuckDB file. They are skipped unless INTERNACIA_TEST_DB_PATH is set, and they fail in CI if it is missing.
• Contract tests (tests/contract/) are the portable subset that internacia-db runs against every fresh build before a release (INTERNACIA_CONTRACT_DB_PATH, falling back to INTERNACIA_TEST_DB_PATH here). They check meta() and the contract, every public method on stable facts, plain Python result types, and that unknown enum values are accepted.

# Unit tests only
pytest

# Download the pinned release database (version and SHA-256 are in tests/internacia-db.json)
python scripts/fetch_test_db.py --version 3.0.0 \
  --sha256 02c33d6822432ef60bce19b81e33b180221f6d6504c3e163f0e14f866254e01d \
  --output .test-data/internacia.duckdb

# Unit and integration tests
INTERNACIA_TEST_DB_PATH=.test-data/internacia.duckdb INTERNACIA_TEST_DB_VERSION=3.0.0 pytest

# With coverage report
pytest --cov=internacia --cov-report=html

New internacia-db releases are picked up by .github/workflows/data-bump.yml: daily, and when internacia-db announces a release, it runs scripts/data_bump.py, which moves tests/internacia-db.json (and TESTED_CONTRACT for a new contract minor) to the new release and opens a pull request. With the same contract major the pull request merges itself once CI passes; a new major waits for SDK changes. It needs the internacia-bot GitHub App, "Allow auto-merge", and branch protection on main that requires the CI checks.

CI runs the full suite on Python 3.11, 3.12 and 3.13 with both the minimum supported duckdb (1.5.4) and the newest allowed one.

Code Formatting

black internacia/ tests/ examples/ scripts/
ruff check internacia/ tests/ examples/ scripts/
mypy internacia/

Pre-commit Hooks

The project includes pre-commit hooks for automated code quality checks:

# Install pre-commit hooks
pre-commit install

# Run hooks manually on all files
pre-commit run --all-files

The hooks automatically check for:

• Code formatting (black)
• Linting (ruff)
• Type checking (mypy)
• Test execution (pytest)
• Trailing whitespace and other code quality issues

Alternative Implementation Approaches for Database Updates

While the current SDK uses a manual update process, here are several alternative approaches that could be implemented to automate database updates:

1. Version Metadata Checking

Approach: Store version/timestamp metadata in the database and provide SDK methods to check for updates.

Implementation:

• Add a metadata table to the database with version, build date, and checksum
• Provide client.get_database_version() and client.check_for_updates() methods
• Compare local version with remote version (from API or file)

Pros: Simple, lightweight, no external dependencies Cons: Requires coordination with internacia-db to maintain version info Complexity: Low

2. CLI Update Command

Approach: Create a CLI tool (e.g., internacia update) that automates the rebuild process.

Implementation:

• Add a CLI entry point using click or argparse
• Command navigates to internacia-db, pulls updates, and rebuilds
• Can be run manually or via cron/scheduled tasks

Pros: User-friendly, automates the manual process Cons: Requires internacia-db to be accessible locally Complexity: Medium

3. Auto-Download from Releases

Approach: Automatically download pre-built database files from GitHub releases.

Implementation:

• Publish database files as GitHub release assets in internacia-db
• SDK checks for new releases and downloads the database file
• Cache downloaded files locally

Pros: No build step required, works for all users Cons: Requires release management, larger repository size Complexity: Medium-High

4. Database Version API

Approach: Check against an API endpoint for the latest database version.

Implementation:

• Create a simple API endpoint (or use internacia-api) that returns latest version info
• SDK queries API on initialization or via a method call
• Compare versions and prompt user to update

Pros: Centralized version management, can provide update notifications Cons: Requires API infrastructure, network dependency Complexity: Medium-High

5. Watch Mode / File Monitoring

Approach: Monitor the database file for changes and reload automatically.

Implementation:

• Use file system watchers (e.g., watchdog library) to detect database file changes
• Automatically reload database connections when file is updated
• Useful for development environments

Pros: Seamless updates during development Cons: File watching overhead, primarily useful for development Complexity: Medium

6. Package Integration with Version Checking

Approach: Include database version checking in SDK initialization with warnings.

Implementation:

• Check database version on client initialization
• Compare against expected/known versions
• Log warnings if database is outdated
• Optionally provide update instructions in the warning

Pros: Proactive user notification, minimal implementation Cons: Requires version metadata in database Complexity: Low-Medium

Recommendation: Start with Version Metadata Checking (#1) combined with Package Integration (#6) for a lightweight solution that provides user feedback. For more advanced use cases, consider CLI Update Command (#2) or Auto-Download from Releases (#3).

License

MIT License - see LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Related Projects

• internacia-db: The data repository
• internacia-api: REST API for internacia-db

Changelog

See CHANGELOG.md for a list of changes.