Metadata-Version: 2.4
Name: moonshiner
Version: 0.4.63
Summary: Seed authoring, trace distillation, and training dataset preparation
Project-URL: Repository, https://github.com/greghavens/moonshiner
Project-URL: Issues, https://github.com/greghavens/moonshiner/issues
Author: Greg Havens
License:                                  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 the 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 the 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 any 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
        
           Copyright 2026 Moonshiner Contributors
        
           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.
License-File: LICENSE
Classifier: License :: OSI Approved :: Apache Software License
Requires-Python: >=3.11
Requires-Dist: huggingface-hub>=0.27
Provides-Extra: huggingface
Requires-Dist: datasets>=3; extra == 'huggingface'
Provides-Extra: parquet
Requires-Dist: pyarrow>=16; extra == 'parquet'
Provides-Extra: trainers
Requires-Dist: transformers>=4.48; extra == 'trainers'
Description-Content-Type: text/markdown

# Moonshiner

Moonshiner authors seeds, creates model traces, judges every trace, retries rejected traces, prepares training data, and publishes accepted trajectories to Hugging Face.

## Install

With pip:

```bash
python3 -m pip install moonshiner
```

With the installer:

```bash
curl -fsSL https://raw.githubusercontent.com/greghavens/moonshiner/main/install.sh | bash
```

## Quick start

After installation, run:

```bash
moonshiner
```

That is the complete quick start. On the first run, Moonshiner confirms where to keep the project's configuration and data, asks for your preferences and anything else it needs, saves the configuration, and starts the pipeline.

Moonshiner asks:

- which harness and model should create traces;
- which harness and model should judge traces;
- which harnesses and models should author and judge seeds, when seed authoring is enabled;
- which provider Pi should use;
- the provider credential, when that provider requires one;
- the Hugging Face dataset target, when publishing is enabled.

Codex and Claude Code can use their existing login sessions. Pi is a harness, so Moonshiner asks which provider Pi should call and configures the matching endpoint, protocol, model, and credential.

After setup, `moonshiner` starts every enabled queue for that project. Seed authoring, tracing, judging, retries, formatting, privacy checks, local append-only storage, and Hugging Face publication continue independently. Active model calls are allowed to finish.

## Check progress

```bash
moonshiner status
```

Status reports active workers, authored seeds, accepted and rejected traces, retraces, rates, session progress, and publishing progress.

Check the installation and project configuration:

```bash
moonshiner doctor
```

## Configure

Run the setup assistant again:

```bash
moonshiner setup
```

Show the current project's configuration:

```bash
moonshiner config show
```

Configure a role directly:

```bash
moonshiner config role trace-author pi PROVIDER/MODEL REASONING
```

```bash
moonshiner config role trace-judge codex gpt-5.6-sol xhigh
```

```bash
moonshiner config role seed-author codex gpt-5.6-sol xhigh
```

```bash
moonshiner config role seed-judge codex gpt-5.6-sol xhigh
```

Save a provider credential:

```bash
moonshiner auth set openrouter
```

Moonshiner stores credentials outside the project directory and removes credentials, user keys, email addresses, and host-identifying paths from data before publication.

## Project storage

Each working directory is a separate Moonshiner project. Configuration and output default to `.moonshiner/` in the current directory.

```bash
moonshiner storage status
```

To use another location, change to that directory and run `moonshiner`. Moonshiner asks you to confirm the directory before creating the project configuration.

## Seeds

Browse the seed catalog:

```bash
moonshiner seeds catalog
```

Search the catalog by category, name, or training tag:

```bash
moonshiner seeds catalog --category parallel-same
```

```bash
moonshiner seeds catalog --name calendar
```

```bash
moonshiner seeds catalog --tag execution:parallel
```

Author and review one seed:

```bash
moonshiner seed run --id calendar-reschedule --brief "Reschedule independent appointments while preserving all constraints" --yes
```

Completed seeds enter the trace queue automatically. Categories and tags organize the catalog and can optionally select training data; they do not create separate trace pipelines.

## Traces

Running `moonshiner` is the normal way to keep tracing continuously. To request a deliberate bounded run:

```bash
moonshiner run --limit 20 --yes
```

Trace workers process separate seeds concurrently. Each seed remains an independent work item: create one trace, judge it, retry it after a judge rejection when attempts remain, then format, scrub, and queue an accepted trace for publication.

Set trace concurrency:

```bash
moonshiner config set pipeline.trace.workers 3
```

Set the maximum attempts for each individual trace:

```bash
moonshiner config set pipeline.trace.max_attempts 2
```

Set Hugging Face publication batch size:

```bash
moonshiner config set publish.batch_size 10
```

These settings are reread between work items. Reducing concurrency does not cancel active model calls.

## Resume existing work

Import an existing directory of traces or prepared rows:

```bash
moonshiner trace import --directory /path/to/existing-data
```

Import an existing Hugging Face dataset:

```bash
moonshiner trace import --hf owner/dataset --revision COMMIT
```

For the configured Hugging Face target, Moonshiner downloads the remote trace file only when the matching local file does not yet exist. Later accepted traces append to the local canonical file. Existing rows are never replaced.

Enable a remote revision check before each append when you need it:

```bash
moonshiner config set publish.check_before_append true
```

## Build and prepare datasets

Build the accepted local traces into a validated dataset:

```bash
moonshiner dataset build
```

Analyze one or more local or revision-pinned Hugging Face datasets before combining them:

```bash
moonshiner dataset analyze --source local:/data/private.jsonl --source hf:HuggingFaceH4/ultrachat_200k@COMMIT#train_sft
```

With no `--source`, Moonshiner analyzes the configured local append-only dataset. A direct Hugging Face file also works without installing dataset tooling:

```bash
moonshiner dataset analyze --source hf-file:owner/dataset@REVISION/path/to/traces.jsonl
```

You can also paste the file's Hugging Face URL directly:

```bash
moonshiner dataset analyze --source https://huggingface.co/datasets/owner/dataset/blob/REVISION/path/to/traces.jsonl
```

The report compares trajectories, rows, target tokens, total tokens, length distributions, categories, tags, sources, multi-turn conversations, direct responses, sequential tool calls, and parallel tool calls. Add `--tokenizer organization/model` for exact tokenizer counts; otherwise Moonshiner clearly labels its token estimate.

Combine local data with revision-pinned Hugging Face datasets:

```bash
moonshiner dataset compose --source local:/data/private.jsonl --source hf:HuggingFaceH4/ultrachat_200k@COMMIT#train_sft --out /data/prepared/train.jsonl
```

Preview a token-budgeted composition without writing it:

```bash
moonshiner dataset compose --source local:/data/private.jsonl --source hf:owner/dataset@COMMIT#train --target-tokens 50000000 --weight-category 'tool-calling=2' --weight-tag 'execution:parallel=3' --tokenizer organization/model --dry-run
```

Remove `--dry-run` to write the composition after reviewing the reported realized mix. Weight rules use `GLOB=WEIGHT`; category, tag, and source-pattern weights require `--target-tokens`. `--weight-unit` selects whether sampling balances rows, target tokens, or total tokens. No curriculum percentage is hard-coded.

Select rows by name, category, or training tag:

```bash
moonshiner dataset compose --source local:/data/all.jsonl --include-category 'tool-*' --include-tag parallel-tool-calls --exclude-tag sensitive --out /data/prepared/selected.jsonl
```

Review training risks:

```bash
moonshiner dataset readiness --source local:/data/prepared/train.jsonl --tokenizer organization/model --context-length 32768
```

Readiness checks context truncation, empty final answers, duplicate prompts, malformed tool sequences, repetitive reasoning, mixed-language scripts, privacy findings, cumulative trajectory prefixes, and small category shares. Analysis and readiness report privacy finding types and counts without printing the affected row contents. They are advisory only; composition and publication remain fail-closed for privacy findings.

Prepare trainer configuration:

```bash
moonshiner dataset prepare --trainer axolotl --input /data/prepared/train.jsonl --model organization/model --tokenizer organization/model --sequence-len 32768 --out /data/prepared/axolotl.json
```

Packing is off unless `--sample-packing` is supplied. Moonshiner normalizes mixed conversation formats, scrubs private data, deduplicates rows, and writes reproducible composition and trainer manifests. Manifests include pinned Hugging Face revisions or local file hashes, tokenizer accounting, filters, requested and realized mixtures, input and output hashes, trainer configuration, package versions, and the exact trainer command.

Accepted traces also receive observed tags such as `response:direct`, `reasoning:planning`, `reasoning:extended`, `reasoning:self-correction`, `reasoning:verification`, `interaction:multi-turn`, `execution:parallel`, and `format:strict-json`. These describe what the trace actually demonstrated; they are catalog and composition metadata, never queue partitions or acceptance gates.

## Hugging Face publishing

Set the target dataset:

```bash
moonshiner config set publish.hf_dataset owner/dataset
```

Log in with the Hugging Face CLI or save the token through Moonshiner. Accepted trajectories are appended locally and published in configured batches. Dataset-card counts and percentages regenerate from the exact published rows.

Manual publication is available for repair or verification:

```bash
moonshiner publish --yes
```

## Seed library

Application releases and seed-catalog releases are versioned separately, so the catalog can grow without requiring an application update.

```bash
moonshiner seeds status
```

```bash
moonshiner seeds update
```

```bash
moonshiner seeds verify
```

Seed identifiers are immutable. Moonshiner never overwrites or removes an existing seed during normal operation.

## Agent use

The repository includes the `skills/moonshiner-runner` skill for agents operating Moonshiner. Agents should use the same installed `moonshiner` commands, project configuration, status output, and released application as human users.

## License

Moonshiner is licensed under the [Apache License 2.0](LICENSE).
