github / github/codeql

getMetrics appears to be miscounting blank lines in docstrings

Open
#20,766 1 comment 0 reactions 0 assignees View on GitHub
Python question
Dominant language
CodeQL
Stars
10.1k
Forks
2.1k
Avg merge
2d 15h
Merged PRs (30d)
141

Description

The definition of the python function `combine_docs` starts on line 227 in [combine_documents/map_reduce.py](https://github.com/langchain-ai/langchain/blob/master/libs/langchain/langchain_classic/chains/combine_documents/map_reduce.py) and ends on line 259.

[func-def.txt](https://github.com/user-attachments/files/23375865/func-def.txt)
Running the above QL script on the latest langchain database produces the values:

func.getMetrics().getNumberOfLines() = 33
func.getMetrics().getNumberOfLinesOfCode() = 27 # Should be 26
func.getMetrics().getNumberOfLinesOfComments() = 2
func.getMetrics().getNumberOfLinesOfDocStrings() = 5

I suspect that the number of blank line in the docstring might be a factor.
The above function contains one blank line in a docstring, while the example below contains two blank lines in a docstring.

The function `collapse_docs` starts on line 67 in [combine_documents/reduce.py](https://github.com/langchain-ai/langchain/blob/master/libs/langchain/langchain_classic/chains/combine_documents/reduce.py) and ends on line 96.

The above QL script produces the values:

func.getMetrics().getNumberOfLines() = 30
func.getMetrics().getNumberOfLinesOfCode() = 12 # Should be 14
func.getMetrics().getNumberOfLinesOfComments() = 0
func.getMetrics().getNumberOfLinesOfDocStrings() = 16

Contributor guide

Open the contributing guide

Research direction

Reproduce the reported metrics for combine_docs in libs/langchain/langchain_classic/chains/combine_documents/map_reduce.py and collapse_docs in combine_documents/reduce.py using the linked QL script and attachment. Compare total, code, comment, and docstring counts with the stated expected values, especially around blank lines in docstrings. Done means the metrics consistently report the expected counts for both examples.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
devtools
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.