Metadata-Version: 2.4
Name: de-agentic
Version: 2.2.0
Summary: Dependency-light CLI for data engineers: profile data, run read-only SQL, read schemas, and ask a local or cloud LLM. Build pipelines with the SDLC mode.
Author: DE Agentic Team
License-Expression: MIT
Project-URL: Homepage, https://github.com/longbuivan/llm-based-data-engineering-agents
Project-URL: Documentation, https://github.com/longbuivan/llm-based-data-engineering-agents/blob/main/docs/CONFIGURATION.md
Project-URL: Repository, https://github.com/longbuivan/llm-based-data-engineering-agents
Project-URL: Issues, https://github.com/longbuivan/llm-based-data-engineering-agents/issues
Project-URL: Changelog, https://github.com/longbuivan/llm-based-data-engineering-agents/releases
Keywords: data-engineering,sql,data-profiling,schema,llm,clickhouse,snowflake,duckdb,sqlite,cli,local-llm,read-only
Classifier: Development Status :: 4 - Beta
Classifier: Environment :: Console
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Information Technology
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Topic :: Database
Classifier: Topic :: Scientific/Engineering :: Information Analysis
Classifier: Topic :: Software Development :: Quality Assurance
Classifier: Topic :: Utilities
Requires-Python: >=3.9
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: pyyaml>=6.0
Requires-Dist: pydantic>=2.0
Provides-Extra: pretty
Requires-Dist: rich>=13.0; extra == "pretty"
Provides-Extra: data
Requires-Dist: duckdb>=0.9; extra == "data"
Requires-Dist: pandas>=2.0; extra == "data"
Provides-Extra: dev
Requires-Dist: pytest>=7.0; extra == "dev"
Requires-Dist: ruff>=0.6; extra == "dev"
Requires-Dist: build>=1.0; extra == "dev"
Requires-Dist: twine>=5.0; extra == "dev"
Provides-Extra: test
Requires-Dist: sqlparse>=0.4; extra == "test"
Requires-Dist: loguru>=0.7; extra == "test"
Requires-Dist: python-dotenv>=1.0; extra == "test"
Provides-Extra: legacy
Requires-Dist: typer>=0.9; extra == "legacy"
Requires-Dist: rich>=13.0; extra == "legacy"
Requires-Dist: duckdb>=0.9; extra == "legacy"
Requires-Dist: pandas>=2.0; extra == "legacy"
Requires-Dist: langchain>=0.2; extra == "legacy"
Requires-Dist: langchain-anthropic>=0.1; extra == "legacy"
Requires-Dist: langchain-openai>=0.1; extra == "legacy"
Requires-Dist: langchain-community>=0.0; extra == "legacy"
Requires-Dist: loguru>=0.7; extra == "legacy"
Requires-Dist: sqlalchemy>=2.0; extra == "legacy"
Requires-Dist: sqlparse>=0.4; extra == "legacy"
Requires-Dist: python-dotenv>=1.0; extra == "legacy"
Dynamic: license-file

# DE Agentic

A dependency-light CLI for data engineers. Profile data files, run read-only
SQL, read schemas, and ask a local or cloud LLM questions about your data.

**Runs on the standard library alone.** No API key, no Docker, no `pip install`
needed to try it.

---

## Quick start

```bash
pip install de-agentic
de init-db
de demo
```

**Thirty seconds, no install:**

```bash
python3 de.py demo
```

## The four commands that matter

```bash
de profile customers.csv        # column types, nulls %, ranges
de query  "SELECT * FROM customers LIMIT 5"  # read-only SQL
de schema                        # tables, columns, relationships
de ask   "how do I find null values?"         # your model, or offline help
```

`profile`, `query`, `schema`, `init-db` never contact a model. `ask` falls
back to local guidance when no model is configured.

## Key features

- **Read-only SQL** — `INSERT`/`UPDATE`/`DELETE`/`DROP` are rejected before any
  connection runs
- **Named database profiles** — configure several databases at once,
  `de dbs` to list them, `--db profile:<name>` to select one
- **SDLC data-pipeline mode** — `de sdlc "daily revenue per customer"` derives
  a full pipeline lifecycle (requirements → design → ingest → transform →
  quality → operate) with checkpoints, mechanical verification, and independent
  QA
- **LLM support** — llama.cpp, Ollama, OpenAI, Anthropic through one
  OpenAI-compatible `/v1/chat/completions` endpoint
- **Local-first** — no API key needed for llama.cpp and Ollama

## Install

```bash
pip install de-agentic
```

Optional extras: `pip install de-agentic[data,legacy]`. See
[QUICKSTART.md](QUICKSTART.md) for the full guide, and
[TROUBLESHOOTING.md](docs/TROUBLESHOOTING.md) if `de doctor` reports issues.

## Setup a model

```bash
de doctor                      # which endpoints are reachable?
de setup                       # detect one and write .env
de ask "which customers have no orders?"
```

## Release history

- **v2.2.0** — Named database profiles, SDLC data-pipeline mode, SQL
  validation, `primary_keys()`
- **v2.1.3** — Config discovery fixed for pip-installed users
- **v2.1.2** — Bug fixes and configuration docs

## License

MIT — see [LICENSE](../LICENSE) for details.
