Skip to content

CHANGELOG.md

Latest commit

 

History

History
42 lines (30 loc) · 5.77 KB

CHANGELOG.md

File metadata and controls

42 lines (30 loc) · 5.77 KB
Raw

Changelog

[v0.13.0] - 2026-07-07

  • Breaking: renamed the starter's placeholder project from my-project / my_project / My Project to piper / Piper. The console command is now uv run piper ..., the template package is piper/, pyproject.toml points at piper.cli:app, and the e2e extract-entities test file no longer carries the project placeholder in its name.
  • Reworked /bootstrap for the single-token placeholder. It now derives distribution, package, and title forms from piper, applies context-aware replacements for command vs import/path positions, edits pyproject.toml by key, and aborts before writing if any placeholder tokens survive.
  • Overhauled the README DevX. The quick start now explains uv run before the first command, shows real expected output, groups the demos around copy-pasteable commands, explains durable vs blocking execution, and adds Mermaid diagrams for hosted-run flow and execution modes.
  • Added demo methods so the starter matches the JS starter. summarize-pdf summarizes a document (PDF) into { title, doc_type, key_points } and demonstrates a file input: piper/file_input.build_document_input() encodes a local file as a base64 data: URL wrapped in a Document envelope. generate-image generates an image from a text prompt and is the slow case that overruns the hosted ~30s blocking cap, making the durable-vs-blocking split concrete. Each is a self-contained "copy me" module under piper/examples/ with a matching main.mthds bundle and its own unit + e2e tests. Ships samples/sample-invoice.pdf to try summarize-pdf on.

[v0.12.0] - 2026-07-06

  • Breaking: dropped Python 3.10 support. pipelex-sdk 0.4.0 no longer supports 3.10, so the starter now requires Python 3.11+ (requires-python = ">=3.11,<3.15"). Removed the Python :: 3.10 classifier and dropped 3.10 from the CI lint/test matrices.
  • Bumped pipelex-sdk to 0.4.0 (pulls in mthds 0.8.1). Re-locked uv.lock and refreshed the dev-tooling pins (ruff, mypy, pyright, pytest).

[v0.11.0] - 2026-07-05

  • Fixed: the demo bundle's list fields (people/orgs/dates) now declare item_type = "text" so the output is typed as list[str], matching the ExtractedEntities model. Without it the runtime built the fields as List[Any].
  • Read a run's output with results.main_stuff. Bumped to pipelex-sdk 0.3.0, which resolves the main output for you on both execution modes: execute returns a PipelexExecuteResult and the durable path a RunResults, and both expose a resolved .main_stuff. The starter's whole output-extraction module (my_project/run_output.py — find_main_content shape-guessing + the to_run_results adapter) is gone; the CLI and the narrower read results.main_stuff directly, and the blocking execute result is adapted onto RunResults inline in the runner. A completed run that delivers no main stuff raises the SDK's MissingMainStuffError instead of yielding None.
  • Breaking: renamed the env var PIPELEX_API_URL to PIPELEX_BASE_URL for consistency with the SDK's base_url naming. There is no read alias — update your .env / environment.
  • Fixed: rewrote the README and CLAUDE.md around the actual my-project CLI (the extract-entities command, the durable/blocking execution modes, and the runs status|result|wait lifecycle). They still described the removed hello_world module, start_and_wait usage, and the find_main_content normalizer, so the quick start's first command errored out for a fresh user.
  • Repository: removed internal-only planning docs (TODOS.md, wip/) that must not ship in a "Use this template" repo.

[v0.10.0] - 2026-07-01

  • Breaking: run methods through the hosted Pipelex API instead of the local pipelex runtime. The pipelex package (and its [tool.uv.sources] git pin) is dropped; the starter now depends on pipelex-sdk (PipelexAPIClient) and python-dotenv.
  • Rewrite my_project/hello_world.py to read the .mthds bundle from disk and run it via client.start_and_wait(pipe_code=..., mthds_contents=[...]), reading the output out of main_stuff / pipe_output.
  • Configuration is now PIPELEX_API_URL / PIPELEX_API_KEY (see .env.example); .env is loaded via python-dotenv.
  • Repoint make validate to plxt lint (offline bundle validation, was pipelex validate --all).
  • Rewrite the test suite to be API-based: offline boot/bundle checks plus API validate (pipelex_api) and a full run (inference). CI no longer runs pipelex init, and make gha-tests / make codex-tests exclude the pipelex_api marker.
  • Prune AWS/doc type-stub dev dependencies that were only needed by the pipelex runtime.
  • Add a /bootstrap skill (.claude/skills/bootstrap/) that turns a fresh clone of this template into a real project. It collects the project name, description, author, repository URL, and license, then renames the package directory and e2e test file, substitutes every placeholder name spelling (dash / underscore / Title Case / CamelCase), applies the chosen license (MIT, proprietary, or another SPDX id) across LICENSE, pyproject.toml, and the README, regenerates uv.lock, and runs the lint/type checks and tests

[v0.9.0] - 2026-06-06

  • Bump pipelex to v0.32.0: See Pipelex changelog here
  • Update tests/integration/test_fundamentals.py to use the new BundleValidator().acquire_and_validate() API (replaces the removed dry_run_pipes / get_library_manager dry-run path)

[v0.8.0] - 2026-05-06

  • Bump pipelex to v0.26.4: See Pipelex changelog here
  • Add pipelex-tools dev dependency
  • Update tests/integration/conftest.py to use the new needs_inference kwarg on Pipelex.make() (replaces deprecated disable_inference)