fix: suppress divide-by-zero RuntimeWarning in BagOfWords idf computation - #89
Merged
Conversation
…tion
idf = np.log(N / ni) can legitimately divide by zero: after stopword
removal (nstopwords > 0), a remaining word can end up with zero
occurrences across every image (ni=0), giving idf=inf for that word.
This is harmless -- every downstream use of idf multiplies it by that
same word's occurrence count, which is also always 0 there, so the
0*inf->nan result gets replaced with 0 before any vector is returned
(already the case at both existing call sites). The equivalent
divide-by-zero one call site down (v = nid / nd * idf) was already
wrapped in np.errstate(divide="ignore", invalid="ignore") -- this one
just wasn't.
Documented the inf possibility in __init__'s docstring, and added a
regression test (existing test_retrieve already exercised the
nstopwords=50 path that triggers this; wrapped it in
warnings.simplefilter("error", RuntimeWarning) so it fails loudly if
this regresses).
Root-caused via RVC3-python's chap12.ipynb (BagOfWords(features, 2_000,
nstopwords=50, seed=0)).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Up to standards ✅🟢 Issues
|
| Metric | Results |
|---|---|
| Complexity | 0 |
| Duplication | 0 |
NEW Get contextual insights on your PRs based on Codacy's metrics, along with PR and Jira context, without leaving GitHub. Enable AI reviewer
TIP This summary will be updated as you push new changes.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
idf = np.log(N / ni)can legitimately divide by zero: after stopword removal (nstopwords > 0), a remaining word can end up with zero occurrences across every image (ni=0), givingidf=inffor that word.idfmultiplies it by that same word's occurrence count, which is also always 0 there, so the0*inf->nanresult gets replaced with 0 before any vector is returned (already the case at both existing call sites). The equivalent divide-by-zero one call site down (v = nid / nd * idf) was already wrapped innp.errstate(divide="ignore", invalid="ignore")-- this one just wasn't.infpossibility in__init__'s docstring.Root-caused via RVC3-python's chap12.ipynb (
BagOfWords(features, 2_000, nstopwords=50, seed=0)).Test plan
test_retrievealready exercised thenstopwords=50path that triggers this; wrapped the relevant call inwarnings.simplefilter("error", RuntimeWarning)so it fails loudly if this regressesRuntimeWarning: divide by zero encountered in divideatBagOfWords.py:142; passes with the fix