fix(document_loaders): ChatGPTLoader misses last message with default num_logs

[leoneperdigao]

Summary

Related Issue

Fixes #465

Issue: Bug: ChatGPTLoader misses the last message when loading default logs
URL: #465

Problem

Checked other resources

• This is a bug, not a usage question.
• I added a clear and descriptive title that summarizes this issue.
• I used the GitHub search to find a similar question and didn't find it.
• I am sure that this is a bug in LangChain Community rather than my code.
• The bug is not resolved by updating to the latest stable version of LangChain Community.
• I read what a minimal reproducible example is (https://stackoverflow.com/help/minimal-reproducible-example).
• I posted a self-contained, minimal, reproducible example. A maintainer can copy it and run it AS IS.

Example Code

'''python
class ChatGPTLoader(BaseLoader):
"""Load conversations from exported ChatGPT data."""

def __init__(self, log_file: str, num_logs: int = -1):
    """Initialize a class object.

    Args:
        log_file: Path to the log file
        num_logs: Number of logs to load. If 0, load all logs.
    """
    self.log_file = log_file
    self.num_logs = num_logs

'''

Error Message and Stack Trace (if applicable)

No response

Description

Description

I found a bug in ChatGPTLoader. The default num_logs in __init__ is set to -1.
In the load method, the slicing [: self.num_logs] is executed when num_logs is not 0.
Since -1 evaluates to True, it performs data[:-1], which excludes the last record of the conversation.

Reproduction Steps

1. Export a single conversation from ChatGPT.
2. Load it using ChatGPTLoader(file_path).
3. The returned document list is empty.

Proposed Fix

Change the default value of num_logs to 0 (which means load all), or handle -1 explicitly.
I have prepared a fix and unit tests for this.

Solution

Tests

Checklist

• Code follows project style guidelines
• Tests added/updated and passing
• Documentation updated (if applicable)
• Self-review completed

Additional Notes