promptdrift-py 0.1.0__tar.gz

promptdrift_py-0.1.0/.gitattributes

@@ -0,0 +1,10 @@
+# Normalize line endings: this repo is tested on Linux, macOS and Windows.
+* text=auto eol=lf
+
+# Windows scripts keep CRLF
+*.ps1 text eol=crlf
+*.bat text eol=crlf
+
+# Binary assets
+*.png binary
+*.gif binary

promptdrift_py-0.1.0/.github/workflows/ci.yml

@@ -0,0 +1,46 @@
+name: CI
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+
+jobs:
+  lint:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: astral-sh/setup-uv@v5
+        with:
+          enable-cache: true
+      - run: uv python install 3.13
+      - run: uv sync
+      - run: uv run ruff check .
+      - run: uv run ruff format --check .
+
+  typecheck:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: astral-sh/setup-uv@v5
+        with:
+          enable-cache: true
+      - run: uv python install 3.13
+      - run: uv sync
+      - run: uv run pyright
+
+  test:
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [ubuntu-latest, macos-latest, windows-latest]
+        python-version: ["3.10", "3.11", "3.12", "3.13"]
+    steps:
+      - uses: actions/checkout@v4
+      - uses: astral-sh/setup-uv@v5
+        with:
+          enable-cache: true
+      - run: uv python install ${{ matrix.python-version }}
+      - run: uv sync --python ${{ matrix.python-version }}
+      - run: uv run pytest --cov=promptdrift --cov-report=term-missing

promptdrift_py-0.1.0/.gitignore

@@ -0,0 +1,31 @@
+# Python
+__pycache__/
+*.py[cod]
+*.egg-info/
+dist/
+build/
+.venv/
+.python-version
+
+# Tooling caches
+.pytest_cache/
+.ruff_cache/
+.coverage
+coverage.xml
+.mypy_cache/
+
+# promptdrift runtime artifacts — baselines are committed, these are not
+.promptest/cache.db*
+.promptest/runs/
+
+# Editors / OS
+.idea/
+.vscode/
+.DS_Store
+Thumbs.db
+
+# promptdrift demo scratch dir
+promptdrift-demo/
+
+# mkdocs build output
+site/

promptdrift_py-0.1.0/CHANGELOG.md

@@ -0,0 +1,34 @@
+# Changelog
+
+All notable changes are documented here. The format follows
+[Keep a Changelog](https://keepachangelog.com/en/1.1.0/) and the project uses
+[Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+## [Unreleased]
+
+### Added
+
+- `judge` assertion: LLM-as-judge rubric scoring on a 1-5 scale. The suite-level `judge:`
+  block configures a separate (usually cheaper) grading model, its scores feed the same
+  multi-sample statistics as deterministic assertions, and unparseable judge replies fail
+  the sample rather than the gate.
+- mkdocs documentation site and a ready-to-copy GitHub Actions workflow for consumer repos.
+
+### Added — M0
+
+- Core domain: `Suite`/`Case` models with readable YAML validation, `{{var}}` templating,
+  and a config fingerprint that guards baseline comparability.
+- Verdict engine: multi-sample pass rates compared with Wilson score intervals; verdicts
+  `regressed` / `improved` / `stable` / `unstable` / `new` / `removed`; gates
+  `--fail-on regression|flaky|any-fail`.
+- Deterministic assertions: `equals`, `contains`, `not_contains`, `regex`, `is_json`,
+  `json_schema`, `latency_under`, `completion_tokens_under`.
+- Providers: `openai_compat` (OpenAI/GLM/DeepSeek/Qwen/Moonshot via `base_url`, retries
+  with backoff) and an offline `mock` (rules, variant cycling for flaky simulation).
+- Response cache (sqlite, read-through) and per-run cost estimation.
+- Baseline snapshots as committed, human-reviewable YAML.
+- CLI: `init`, `run`, `diff`, `approve`, `show`, `cost`, `demo`, `schema`, `version`;
+  JUnit XML and markdown report output; exit codes 0 pass / 1 gate failure / 2 error.
+- Published JSON Schema for suite files (editor completion).
+- Offline test suite, ruff and pyright clean, CI matrix Python 3.10-3.13 on
+  Linux/macOS/Windows.

promptdrift_py-0.1.0/CONTRIBUTING.md

@@ -0,0 +1,64 @@
+# Contributing
+
+Issues and PRs are welcome. This file covers setup, how the codebase is organized, and the
+rules that keep the tool trustworthy.
+
+## Setup
+
+```bash
+git clone https://github.com/ritaprieto900/promptdrift
+cd promptdrift
+uv sync          # installs the dev group: pytest, ruff, pyright, respx, ...
+uv run pytest    # must be green before any PR
+```
+
+CI enforces the same gates you should run locally:
+
+```bash
+uv run ruff check .               # lint
+uv run ruff format .              # formatting
+uv run pyright                    # types, zero errors
+uv run pytest --cov=promptdrift   # coverage gate at 90%
+```
+
+## Code layout
+
+Dependencies point downward only. The domain layer imports nothing that does I/O.
+
+```
+CLI (cli/app.py)        entry points, exit-code mapping
+Services (services/)    orchestration: Runner, Differ, Reporter, Cost, Loader, Judge
+Domain (domain/)        pure: suite, assertions, results, snapshot, verdict engine
+Ports (ports.py)        Provider / Cache protocols
+Adapters (adapters/)    mock, openai_compat, sqlite cache, YAML snapshot store
+```
+
+Two consequences worth knowing before you change things:
+
+- Every verdict-path test runs offline on the mock provider. Adapter HTTP behavior is
+  simulated with respx, so CI never touches a real endpoint.
+- Report formats are pinned by golden files in `tests/golden/`. If you change a format on
+  purpose, regenerate with `PROMPTDRIFT_REGEN_GOLDEN=1 uv run pytest tests/golden -q`.
+
+The JSON Schema in `schema/` is generated from the pydantic models. After touching any
+Suite/Case/provider/assertion model, run `uv run python scripts/gen_schema.py`;
+`tests/test_schema_sync.py` fails if you forget.
+
+## Design rules
+
+These are product decisions, and PRs that weaken them need to argue why in the description.
+
+1. Noise alone can never fail the default gate. A regression verdict requires Wilson
+   intervals to separate, or an all-pass to all-fail flip.
+2. Prompt content changes are measured, never treated as staleness. Only model, sampling
+   params, sample count, and judge model feed the config fingerprint.
+3. Exit codes are a contract: 0 pass, 1 gate or assertion failure, 2 config or runtime error.
+4. The demo must work with zero API keys and zero network.
+5. Baselines are committed, human-reviewable YAML. No binary or opaque state.
+
+## Submitting
+
+Use Conventional Commits (`feat:`, `fix:`, `docs:`, `test:`, `refactor:`), one logical
+change per PR, tests included for behavior changes. Good first contributions are usually
+the smaller roadmap items in the README; ask in an issue before picking up anything that
+touches the verdict engine.

promptdrift_py-0.1.0/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of whether to use or redistribute the Work and any
+      output generated pursuant to this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

promptdrift_py-0.1.0/PKG-INFO

@@ -0,0 +1,215 @@
+Metadata-Version: 2.5
+Name: promptdrift-py
+Version: 0.1.0
+Summary: Snapshot-diff-first prompt regression testing with flake-aware verdicts and CI gating. Know what your prompt change actually did.
+Project-URL: Homepage, https://github.com/ritaprieto900/promptdrift
+Project-URL: Issues, https://github.com/ritaprieto900/promptdrift/issues
+Project-URL: Changelog, https://github.com/ritaprieto900/promptdrift/blob/main/CHANGELOG.md
+Author: promptdrift contributors
+License-Expression: Apache-2.0
+License-File: LICENSE
+Keywords: ai,ci,eval,llm,prompt,regression,snapshot,testing
+Classifier: Development Status :: 4 - Beta
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Developers
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Software Development :: Quality Assurance
+Classifier: Topic :: Software Development :: Testing
+Classifier: Typing :: Typed
+Requires-Python: >=3.10
+Requires-Dist: httpx>=0.27
+Requires-Dist: jsonschema>=4.21
+Requires-Dist: pydantic<3,>=2.7
+Requires-Dist: pyyaml>=6.0
+Requires-Dist: rich>=13.7
+Requires-Dist: typer>=0.12
+Description-Content-Type: text/markdown
+
+# promptdrift
+
+**Know what your prompt change actually did.**
+
+Prompt regression testing with CI gating. Snapshot-diff first, flake-aware by design, Python native.
+
+[![CI](https://github.com/ritaprieto900/promptdrift/actions/workflows/ci.yml/badge.svg)](https://github.com/ritaprieto900/promptdrift/actions/workflows/ci.yml)
+[![Python](https://img.shields.io/badge/python-3.10%E2%80%933.13-blue)](https://pypi.org/project/promptdrift/)
+[![License](https://img.shields.io/badge/license-Apache--2.0-green)](LICENSE)
+[![Code style: ruff](https://img.shields.io/badge/code%20style-ruff-261230)](https://github.com/astral-sh/ruff)
+
+[中文文档](README.zh-CN.md)
+
+Teams that ship LLM features edit prompts constantly, and every edit can quietly break
+behavior that used to work. promptdrift catches that: it re-runs the same cases against your
+prompt, compares pass rates against a baseline snapshot you commit to git, and fails CI when
+behavior got statistically worse. Noise never trips the gate.
+
+## Quick tour, no API key needed
+
+```bash
+pipx install promptdrift
+promptdrift demo
+```
+
+The demo runs a complete loop offline on a scripted mock provider:
+
+1. `run` executes a suite (3 samples per case)
+2. `approve` records the behavior as the baseline
+3. the mock answer gets "improved" and loses its structured steps
+4. `diff` catches the drift and exits 1, the way CI would
+
+```text
+step 4/4 — diff: catch the drift
+┌──────────────┬─────────────────────┬────────┬───────┬────────────────┐
+│ case         │ assertion           │ before │ after │ verdict        │
+├──────────────┼─────────────────────┼────────┼───────┼────────────────┤
+│ refund-steps │ contains-申请退款    │ 3/3    │ 0/3   │ 🔴 regressed   │
+│              │ regex-3-5-个工作日   │ 3/3    │ 0/3   │ 🔴 regressed   │
+│ invoice-json │ is_json-yes         │ 3/3    │ 3/3   │ ✅ stable      │
+└──────────────┴─────────────────────┴────────┴───────┴────────────────┘
+❌ gate FAILED — 1 regressed case(s)
+```
+
+## Why plain assertions aren't enough
+
+LLM output is non-deterministic. A CI check like "output must contain X" fails randomly,
+gets retried, and eventually gets disabled. The verdict engine handles the noise instead:
+
+| Verdict | Meaning | Gate behavior |
+|---|---|---|
+| 🔴 `regressed` | pass rate fell beyond sampling noise (Wilson score intervals separate) | fails `regression` (default) |
+| ⚠️ `unstable` | samples disagree where the baseline was deterministic | fails only `flaky` |
+| 🟢 `improved` | pass rate rose beyond noise | never fails |
+| ✅ `stable` | within noise | passes |
+| 🆕 `new` / ➖ `removed` | case or assertion added or removed since the baseline | fails `any-fail` when imperfect |
+
+Concretely: at `samples: 3`, a 3/3 → 2/3 drop is ⚠️ unstable, a warning. At `samples: 10`,
+10/10 → 3/10 is 🔴 regressed. Raise the sample count to sharpen the statistics; run
+`promptdrift cost` first to see what that costs.
+
+## How it fits together
+
+```
+promptest.yaml ──► Runner ──► Run ──► Differ ◄── Snapshot (.promptest/baselines/*.snap.yaml)
+                       │                          ▲
+                       └── provider ──► sqlite cache
+```
+
+Baselines are YAML files you commit. When a PR touches one, that diff *is* the behavior
+change under review, the same way jest snapshots work. A config fingerprint (model, sampling
+params, sample count, judge model) protects comparability: swap the model and the old
+baseline goes stale instead of being silently mis-compared. Editing prompt text, vars, or
+mock answers is never staleness; that is exactly the change the diff should measure.
+
+One provider adapter covers every OpenAI-compatible endpoint (OpenAI, GLM, DeepSeek, Qwen,
+Moonshot, vLLM, Ollama), configured with `base_url`.
+
+## Suite file
+
+```yaml
+$schema: https://raw.githubusercontent.com/ritaprieto900/promptdrift/main/schema/promptest.schema.json
+suite: support-agent
+provider:
+  openai_compat:
+    model: glm-4.7
+    base_url: https://open.bigmodel.cn/api/paas/v4
+    api_key_env: ZHIPUAI_API_KEY
+    pricing: {prompt_per_1m: 1.0, completion_per_1m: 8.0}
+judge:                          # optional: grades `judge` assertions
+  openai_compat:
+    model: glm-4.7-flash
+    base_url: https://open.bigmodel.cn/api/paas/v4
+    api_key_env: ZHIPUAI_API_KEY
+    temperature: 0.1
+samples: 3
+cases:
+  - id: refund-steps
+    messages:
+      - role: user
+        content: "订单有点问题，我想申请{{topic}}"
+    vars:
+      topic: 退款
+    assertions:
+      - contains: "申请退款"
+      - regex: "3-5 个工作日"
+      - not_contains: "抱歉"
+      - latency_under: 3000
+      - judge:
+          rubric: |
+            回复是否包含明确的退款步骤（入口、操作、时限）？
+            5 = 步骤完整且语气友好；1 = 没有回答问题。
+          min_score: 4
+```
+
+Deterministic assertions: `equals`, `contains`, `not_contains`, `regex`, `is_json`,
+`json_schema`, `latency_under`, `completion_tokens_under`. The `judge` assertion sends each
+sample to the judge model with your rubric and expects a 1-5 score; `min_score` sets the
+pass bar. Judges are only human (well, only models), so their scores feed the same
+multi-sample statistics as everything else: a wavering judge shows up as ⚠️ unstable, not as
+a blocked merge.
+
+## Commands
+
+| Command | Purpose |
+|---|---|
+| `promptdrift init` | scaffold a starter suite (mock provider, runs offline) |
+| `promptdrift run` | execute once; exit 1 if any assertion failed |
+| `promptdrift approve` | record current behavior as the baseline |
+| `promptdrift diff` | run and compare against the baseline; this is the CI gate |
+| `promptdrift show` | inspect the latest run or the baseline |
+| `promptdrift cost` | estimate calls and cost (when pricing is configured) |
+| `promptdrift demo` | the offline end-to-end walkthrough |
+| `promptdrift schema` | JSON Schema for suite files (editor completion) |
+
+`diff` options that matter in CI: `--fail-on regression|flaky|any-fail`,
+`--require-baseline`, and `--md report.md` to write a PR-ready markdown report.
+
+Exit codes: 0 pass, 1 gate or assertion failure, 2 config or runtime error.
+
+```yaml
+# GitHub Actions, minimal version (see docs/ci.md for the full one)
+- run: uv tool install git+https://github.com/ritaprieto900/promptdrift
+- run: promptdrift diff --require-baseline --md pr-report.md
+  env:
+    OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
+```
+
+## How it compares
+
+| | promptdrift | promptfoo | DeepEval |
+|---|---|---|---|
+| Language | Python | Node/TypeScript | Python |
+| Core primitive | baseline snapshot diff | eval runs + web viewer | pytest metrics |
+| Non-determinism | multi-sample + interval statistics | per-run assertions | metric thresholds |
+| Review flow | snapshot diff in the PR | open the web UI | read pytest output |
+
+promptfoo is the broader eval platform and DeepEval is strong on metric research. This
+project stays narrow: gate prompt changes in CI, with a verdict you can trust.
+
+## Roadmap
+
+- selective re-runs of changed cases only; embedding-similarity assertion
+- composite GitHub Action + sticky PR comments once the PyPI release lands
+- multi-suite projects, cross-model comparison mode
+
+## Development
+
+```bash
+git clone https://github.com/ritaprieto900/promptdrift
+cd promptdrift
+uv sync
+uv run pytest              # 150+ tests, fully offline
+uv run ruff check .
+uv run pyright
+```
+
+The whole test suite runs on the mock provider, so CI never touches a real endpoint.
+CONTRIBUTING.md has the architecture map and the design invariants.
+
+## License
+
+[Apache-2.0](LICENSE)