astro-imaging-agent 0.0.0__tar.gz

astro_imaging_agent-0.0.0/AGENTS.md

@@ -0,0 +1,12 @@
+# Project instructions
+
+- Never create a Git commit or push changes unless the user explicitly changes this instruction.
+- Prepare completed changes in the Git staging area. Preserve unrelated existing edits.
+- Use Poetry for dependency management and execution; do not introduce uv.
+- Write meaningful tests for new behavior and bug fixes, using small synthetic images.
+- Run pytest, Ruff formatting/linting, mypy, and package build checks before handing off work.
+- Keep image tools, analysis, pipeline execution, CLI, and planning independent.
+- A saved PipelineDefinition must execute without an agent or an LLM provider.
+- Processing must be deterministic and must preserve relevant FITS metadata.
+- Document public APIs and behavior, including numerical limitations.
+- Do not add GPU processing, external LLM integrations, or unrelated dependencies without a task requiring them.

astro_imaging_agent-0.0.0/LICENSE

@@ -0,0 +1,202 @@
+
+                                 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 using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under 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.

astro_imaging_agent-0.0.0/NOTICE

@@ -0,0 +1,7 @@
+astro-imaging-agent
+Copyright 2026 Alpha Codes s.r.o.
+
+Made by Alpha Codes s.r.o.
+Author and maintainer: Ondřej Brůha <ondrej.bruha@alphacodes.eu>
+
+Licensed under the Apache License, Version 2.0.

astro_imaging_agent-0.0.0/PKG-INFO

@@ -0,0 +1,316 @@
+Metadata-Version: 2.4
+Name: astro-imaging-agent
+Version: 0.0.0
+Summary: Deterministic astronomical FITS processing with reproducible pipelines and provider-independent planning
+License-Expression: Apache-2.0
+License-File: LICENSE
+License-File: NOTICE
+Keywords: astronomy,fits,image-processing,cli,pipelines
+Author: Ondřej Brůha
+Author-email: ondrej.bruha@alphacodes.eu
+Maintainer: Ondřej Brůha
+Maintainer-email: ondrej.bruha@alphacodes.eu
+Requires-Python: >=3.12,<4.0
+Classifier: Development Status :: 3 - Alpha
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Science/Research
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
+Classifier: Topic :: Scientific/Engineering :: Astronomy
+Classifier: Typing :: Typed
+Provides-Extra: anthropic
+Provides-Extra: gemini
+Provides-Extra: llm
+Provides-Extra: openai
+Requires-Dist: Pillow (>=11.0,<13.0)
+Requires-Dist: PyYAML (>=6.0.2,<7.0)
+Requires-Dist: anthropic (>=0.60,<1.0) ; extra == "anthropic"
+Requires-Dist: anthropic (>=0.60,<1.0) ; extra == "llm"
+Requires-Dist: astropy (>=7.0,<9.0)
+Requires-Dist: google-genai (>=1.30,<3.0) ; extra == "gemini"
+Requires-Dist: google-genai (>=1.30,<3.0) ; extra == "llm"
+Requires-Dist: numpy (>=2.0,<3.0)
+Requires-Dist: openai (>=1.90,<3.0) ; extra == "llm"
+Requires-Dist: openai (>=1.90,<3.0) ; extra == "openai"
+Requires-Dist: photutils (>=2.2,<4.0)
+Requires-Dist: pydantic (>=2.10,<3.0)
+Requires-Dist: pypng (>=0.20220715.0,<1.0)
+Requires-Dist: scipy (>=1.14,<2.0)
+Requires-Dist: tifffile (>=2025.1,<2027.0)
+Requires-Dist: typer (>=0.19,<1.0)
+Description-Content-Type: text/markdown
+
+# astro-imaging-agent
+
+**Made by Alpha Codes s.r.o.** Author and maintainer: **Ondřej Brůha**
+([ondrej.bruha@alphacodes.eu](mailto:ondrej.bruha@alphacodes.eu)).
+
+`aia` is a Python CLI and modular astronomical image-processing library. It measures
+images, applies deterministic processing tools, and saves replayable YAML pipelines
+with JSON processing reports. The default rule-based planner runs entirely locally.
+Optional OpenAI, Anthropic and Google Gemini planners interpret natural-language goals.
+
+The purpose of the LLM integration is **planning and orchestration**.
+A generative model never modifies pixels. The processing engine accepts a validated
+pipeline and remains usable independently of any agent.
+
+## Installation
+
+Requires Python **3.12+**. Package name: `astro-imaging-agent`; Python import:
+`astroagent`; CLI command: **`aia`**.
+
+After the first PyPI release:
+
+```bash
+python -m pip install astro-imaging-agent
+aia --help
+```
+
+For development from source, see [development.md](development.md). The core runs
+locally; optional LLM planners require their SDK extra and API key.
+
+## Supported image formats
+
+| Format | Reading | Writing | Precision and metadata |
+| --- | --- | --- | --- |
+| FITS (`.fit`, `.fits`, `.fts`, also `.gz`) | Mono and unambiguous RGB image HDUs | Mono/RGB | Integer or float; scientific header, WCS, COMMENT and HISTORY preserved; structural, scaling and checksum cards regenerated |
+| TIFF (`.tif`, `.tiff`) | Single mono/RGB image; contiguous or planar RGB | Mono/RGB | Integer or float arrays preserved; own exports embed original metadata and the complete FITS header in ImageDescription |
+| PNG | Mono/RGB, including **16-bit RGB** and opaque palettes | Mono/RGB | uint8 stays 8-bit; uint16/normalized floats export as 16-bit; lossless compression, float quantization |
+| JPEG (`.jpg`, `.jpeg`) | Mono/RGB | Mono/RGB | 8-bit **lossy**, quality 95 with no chroma subsampling |
+| WebP | Single mono/RGB frame | 8-bit RGB | Lossless compression; floats/uint16 quantized to 8-bit; mono exports decode as RGB |
+| BMP | Mono/RGB | Mono/RGB | 8-bit, lossless compression; floats/uint16 quantized to 8-bit |
+
+PNG/JPEG/WebP/BMP export accepts float pixels in **0..1**, or unsigned 8/16-bit
+integer input scaled by its dtype range. It never silently stretches raw scientific
+data. Add `normalize` or `stretch` before exporting an unbounded float image.
+TIFF and FITS preserve unbounded floats and negative sky-subtracted residuals.
+Standard raster exports retain original metadata in the processing JSON sidecar;
+they do not embed a FITS header. JPEG/WebP/BMP are display products, not scientific
+intermediates. Alpha channels, animated images, TIFF stacks and non-RGB FITS cubes
+are rejected with a readable error. Ordinary JPEG/WebP/BMP EXIF orientation is applied
+when loading. Existing tone mapping in display images cannot be reliably inferred.
+
+FITS RGB is internally `(height, width, 3)`. A cube without an `ASTRCHAX` card
+must have exactly one end axis of length three; a cube with both end axes of length
+three is ambiguous. Own exports declare the NumPy storage channel axis in
+`ASTRCHAX`. Loaded FITS storage orientation is retained, preserving WCS axis mapping.
+
+## CLI examples
+
+```bash
+aia inspect image.fit
+aia inspect image.fit --json
+aia analyze-background image.fit
+aia analyze-stars image.fit --fwhm 3 --threshold-sigma 5
+
+aia normalize image.fit normalized.fit
+aia stretch image.fit stretched.fit --method asinh --strength 0.6
+aia denoise image.fit smoothed.fit --sigma 0.8
+aia background-extract image.fit corrected.fit --grid-size 8 --polynomial-degree 2
+
+# The output extension selects the format.
+aia stretch image.fit preview.png --strength 0.6
+aia normalize photo.tiff preview.jpg
+aia inspect preview.png --json
+
+aia run examples/pipeline.yaml --input image.fit --output processed.fit
+aia agent image.fit "zpracuj tento snímek přirozeně a nepřepal hvězdy" --dry-run --explain
+aia agent image.fit "zpracuj tento obrázek" --output processed.fit --explain
+
+aia tools
+aia -v run examples/pipeline.yaml --input image.fit --output processed.fit
+aia -vv inspect missing.fit
+```
+
+Verbosity flags precede the command: `-v` logs steps, `-vv` includes debug tracebacks.
+Expected errors show `Error: ...` with exit code 1; normal execution has no debug logs.
+Single-tool commands can omit the output, defaulting to `<input-stem>.processed.fit`.
+Use `--overwrite` to replace output artifacts; input replacement is always refused.
+`python -m astroagent` provides the same CLI.
+
+## Pipeline and processing reports
+
+```yaml
+version: 1
+steps:
+  - tool: background_extract
+    params:
+      grid_size: 8
+      polynomial_degree: 2
+      sigma_clipping_threshold: 3.0
+  - tool: denoise
+    params:
+      method: gaussian
+      sigma: 0.8
+  - tool: stretch
+    params:
+      method: asinh
+      strength: 0.6
+```
+
+Every processing command saves:
+
+```text
+processed.fit
+processed.pipeline.yaml
+processed.processing.json
+```
+
+The saved YAML contains all **resolved defaults**, so changes in future defaults do
+not change the plan. Replay it with `aia run processed.pipeline.yaml --input image.fit
+--output replay.fit`. All steps and parameters are validated before processing starts;
+unknown tools, misspelled fields, unsupported versions and invalid values are rejected.
+The executor stops at the first failure and identifies its step number.
+
+The JSON report contains input/output paths, package/dependency versions, pixel
+SHA-256 hashes, resolved pipeline, per-step before/after metrics, elapsed milliseconds,
+warnings and (for agent runs) the request, short explanations, and full analysis of
+the initial/final images. Pixel hashes canonicalize data as little-endian float64
+with shape; the output-file hash identifies the encoded artifact. Raster export
+quantization is also measured separately. Timings are informational and vary between
+runs. Replay equality requires the same inputs, tool versions, parameters and numerical
+environment; cross-platform floating-point bitwise identity is not promised.
+
+Outputs are checked for collisions before execution. Each artifact is written
+atomically, but the three-file bundle is not a transaction: a disk failure may leave
+a partial bundle. Intermediate files are not saved.
+
+## Architecture
+
+```text
+src/astroagent/
+  models/       ndarray-based AstroImage and shared strict schema model
+  io/           FITS/raster adapters and metadata handling
+  analysis/     finite statistics, robust background, photutils stars, quality
+  tools/        typed deterministic tools and immutable per-executor registry
+  pipeline/     Pydantic definitions, YAML/JSON, executor and execution history
+  agent/        planner protocols, rules, and analysis/plan/execute orchestration
+  cli/          Typer argument parsing, presentation and error boundary
+```
+
+The central boundary is:
+
+```text
+Planner -> PlanResult(PipelineDefinition, explanations)
+       -> PipelineExecutor -> ImageTool -> ToolResult
+```
+
+`PipelineExecutor` does not import the agent layer. `AstroImage` is a dataclass so
+large NumPy arrays do not pass through Pydantic. Schemas, metrics, plans and reports
+use Pydantic v2. Tools operate in float64, return new arrays, copy metadata/headers,
+and provide before/after metrics and warnings. There is no global mutable registry.
+SciPy supplies Gaussian filtering and NumPy supplies polynomial fitting; scikit-image
+is not needed for this implementation. Optional future backends (OpenCV, CuPy,
+PyTorch, SEP, astroalign, ccdproc) can live behind new tools/adapters.
+
+## Analysis and numerical behavior
+
+- `inspect` reports dimensions `(height, width)`, channels, dtype, finite min/max,
+  mean/median/std, percentiles 1/5/50/95/99, saturation and JSON-safe metadata.
+  RGB summary values aggregate channel samples. NaN/Inf are excluded with a warning;
+  processing rejects them until a future repair/masking tool is provided.
+- Saturation uses `SATURATE`, then the integer dtype ceiling, or 1 for bounded float
+  images. Otherwise its fraction is `null` with a warning. It counts **channel samples**,
+  not whole RGB pixels. A normalized maximum is not evidence of physical detector
+  saturation; saturation fractions after stretching are display-range measurements.
+- Background uses sigma-clipped tile medians and a fitted plane. `sigma` is clipped
+  residual noise; `gradient_estimate` is the fitted surface's 95th–5th percentile
+  range in input units. RGB analysis uses equal-weight channel averaging.
+- Stars use photutils `DAOStarFinder`. Positive local second moments give approximate
+  Gaussian FWHM in pixels, ellipticity `1-b/a`, and eccentricity `sqrt(1-(b/a)^2)`.
+  Failed detection or unreliable shapes return `null` and warnings. This is a rough
+  quality indicator, not precision PSF fitting or calibrated photometry.
+- Normalize uses one global min/max scale across RGB; constant images map to the
+  lower endpoint. The target interval must satisfy `0 <= lower < upper <= 1`.
+- Stretch also uses global endpoints. `black_point` is in input units and must be
+  below the maximum. `linear` ignores strength; `asinh` uses
+  `gain = 10**(3*strength)-1` and `asinh(gain*x)/asinh(gain)`, with strength zero
+  reducing to linear. Strength is in 0..1. Explicit black points can clip shadows.
+- Gaussian denoising uses reflected spatial edges and never mixes RGB channels.
+  It softens stars; the report makes that limitation explicit.
+- Background extraction subtracts a total-degree 0–3 polynomial from each channel
+  independently. It removes offset as well as gradient and preserves negative
+  values. Extended nebulosity, gradients within tiles and crowded fields can bias
+  automatic samples; rank-deficient grids are rejected. It is not a full DBE tool.
+
+## Agent flow
+
+`AgentExecutor` loads an image, inspects it, estimates sky and stars, then sends only
+`ImageMetrics` and `ToolDescription` schemas to `Planner.create_plan`. The CLI prints
+the pipeline, executes it through the ordinary executor, reanalyzes, and saves all
+artifacts. `--dry-run` displays the plan without writing any files. `--explain`
+displays concise explicit reasons; no hidden model reasoning is stored.
+
+`RuleBasedPlanner` compares gradient/noise against the robust 99th–1st percentile
+range: gradient over 15% selects a degree-1 background model, noise over 8% selects
+mild Gaussian denoising, and a dark median suggests an asinh stretch. `ASTRSTR`
+marks images stretched by this tool. Out-of-range non-linear images are normalized.
+Constant images have empty plans. Only advertised tools can be selected.
+
+**The rule-based MVP does not understand the free-text request.** It records it for
+provenance and plans from measurements. Its thresholds and linearity inference are
+heuristics, so review `--dry-run --explain` before relying on them. `LLMPlanner` provides the common boundary for `OpenAIPlanner`, `AnthropicPlanner`
+and `GeminiPlanner`; each returns the same validated `PlanResult` and has no access
+to image arrays or engine internals.
+
+## LLM planners
+
+Install only the provider you need, or all three:
+
+```bash
+python -m pip install 'astro-imaging-agent[openai]'
+python -m pip install 'astro-imaging-agent[anthropic]'
+python -m pip install 'astro-imaging-agent[gemini]'
+python -m pip install 'astro-imaging-agent[llm]'
+```
+
+Set `OPENAI_API_KEY`, `ANTHROPIC_API_KEY`, or `GEMINI_API_KEY` in your environment.
+Choose a model ID available to your account explicitly; no model is hardcoded:
+
+```bash
+aia agent image.fit "preserve stars and bring out faint detail" --provider openai --model YOUR_GPT_MODEL --dry-run --explain
+aia agent image.fit "natural processing" --provider anthropic --model YOUR_CLAUDE_MODEL --output processed.fit
+aia agent image.fit "remove the sky gradient" --provider gemini --model YOUR_GEMINI_MODEL --dry-run
+```
+
+`--provider rules` is the default. `--timeout` controls SDK request timeout (default
+60 seconds). LLM dry-run calls the selected provider to generate a plan but never
+writes or processes an image. The provider receives the request, image measurements,
+metadata and tool schemas; image pixels/files and API keys are not included in the
+planning payload. The report records provider and model, while the saved YAML replays
+offline with `aia run`. Planning itself can vary between API calls.
+
+OpenAI uses [Responses JSON mode](https://developers.openai.com/api/docs/guides/structured-outputs),
+Claude uses a [forced plan-submission tool call](https://platform.claude.com/docs/en/api/python/messages/create),
+and Gemini uses [Google GenAI JSON generation](https://github.com/googleapis/python-genai).
+Extensible parameter dictionaries are checked locally with Pydantic and the registry;
+provider JSON mode alone does not guarantee schema validity. Malformed, incomplete,
+unknown-tool or invalid-parameter plans are rejected before processing. There is no
+automatic fallback to rules when an API call fails. API errors omit SDK response
+bodies; keys are read from the environment and are never written to artifacts.
+Tests use fake clients and do not make paid API calls.
+
+## Contributing
+
+See [development.md](development.md) for Poetry setup, test commands, adding tools,
+versioning, and PyPI/GitHub release configuration.
+
+## MVP scope and next steps
+
+Implemented: multi-format I/O, inspection, background/star analysis, normalization,
+linear/asinh stretching, Gaussian denoising, polynomial sky subtraction, tool schemas,
+YAML replay, processing history, rule-based orchestration and packaging automation.
+
+Not included: stacking, bias/dark/flat calibration, debayering, plate solving,
+photometric color calibration, neural processing, GPU acceleration or distribution.
+The next useful step is explicit sky/star masks and stronger quality checks for
+crowded fields and nebulosity, followed by dataset-based evaluation of LLM-generated
+plans. Scientific FITS/TIFF artifacts
+should remain the processing intermediates.
+
+## License
+
+Apache License 2.0. See [LICENSE](LICENSE) and [NOTICE](NOTICE).
+