fix(data): pass per-sample tools list for preformatted batches

[curnane-lab]

Motivation

preprocess_conversations defaults tools=[[]], expecting a list-of-lists with one entry per conversation so that zip(conversations, tools) pairs each conversation with its tool list.

In the is_preformatted=True branch of preprocess_function, however, the code previously passed a single empty list [] instead of a per-sample list. This caused zip(conversations, tools) to stop after the first sample, silently dropping every remaining sample in the batch. As a result, preformatted datasets were effectively truncated to one sample per batch during training data preparation.

This PR fixes the mismatch and keeps the preformatted path consistent with the non-preformatted path.

Modifications

• specforge/data/preprocessing.py
  • In the is_preformatted branch of preprocess_function, pass tools=[[] for _ in range(len(examples["text"]))] so that every sample in the batch gets its own (empty) tools entry.
  • No changes to the non-preformatted path or to preprocess_conversations itself.

Related Issues

Fixes #555

Accuracy Test

N/A — this is a data-preprocessing bug fix; no model architecture or kernel code is changed.

Benchmark & Profiling

N/A — the change only restores the intended batch size for preformatted datasets. No throughput or latency impact is expected.

Checklist

• Format your code according to the Code Formatting with Pre-Commit.
• Add unit tests as outlined in the Running Unit Tests.
• Update documentation / docstrings / example tutorials as needed, according to Writing Documentation.
• Provide throughput / latency benchmark results and accuracy evaluation results as needed, according to Benchmark and Profiling and Accuracy Results.
• For reviewers: If you haven't made any contributions to this PR and are only assisting with merging the main branch, please remove yourself as a co-author when merging the PR.
• Please feel free to join our Slack channel at https://sgl-fru7574.slack.com/archives/C09784E3EN6 to discuss your PR.

[gemini-code-assist]

Code Review

This pull request updates the preprocessing function in specforge/data/preprocessing.py to pass an empty list of tools for each text example when handling preformatted data. I have no feedback to provide.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.