fix(search): ignore invalid photo exposure times

[fredrikblau]

Summary

Apache Tika can report exif:ExposureTime as 0.0 for embedded placeholder images. Dividing by that value creates +Inf, which later makes the Graph photo JSON impossible to marshal and prevents the containing file from being indexed.

Ignore non-positive and non-finite exposure values while preserving valid photo metadata.

Testing

• gofmt -w services/search/pkg/content/tika.go services/search/pkg/content/tika_test.go
• go test ./services/search/pkg/content
• git diff --check

Fixes #3267

[codacy-production]

Up to standards ✅

🟢 Issues 0 issues

Results:
0 new issues

View in Codacy

🟢 Metrics 1 duplication

Metric	Results
Duplication	1

View in Codacy

🟢 Coverage 100.00% diff coverage · +0.00% coverage variation

Metric	Results
Coverage variation	✅ +0.00% coverage variation (-1.00%)
Diff coverage	✅ 100.00% diff coverage

View coverage diff in Codacy

Coverage variation details

	Coverable lines	Covered lines	Coverage
Common ancestor commit (f60e2e5)	87708	20427	23.29%
Head commit (67124a6)	87709 (+1)	20430 (+3)	23.29% (+0.00%)

Coverage variation is the difference between the coverage for the head and common ancestor commits of the pull request branch: <coverage of head commit> - <coverage of common ancestor commit>

Diff coverage details

	Coverable lines	Covered lines	Diff coverage
Pull request (#3420)	6	6	100.00%

Diff coverage is the percentage of lines that are covered by tests out of the coverable lines that the pull request added or modified: <covered lines added or modified>/<coverable lines added or modified> * 100%

_{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.}

[butonic]

@dschmidt can I get your review on this?

[dschmidt]

Thanks, good catch!

One suggestion: I'd guard the computed denominator instead of the input and just skip the block instead of returning early. ExposureTime happens to be the last field today, but if someone adds another field below, the early return would silently drop it. Checking the result also covers the subnormal case where i > 0 but 1/i still overflows to +Inf:

if v, err := getFirstValue(meta, "exif:ExposureTime"); err == nil {
	if i, err := strconv.ParseFloat(v, 64); err == nil {
		if d := math.Round(1 / i); !math.IsNaN(d) && !math.IsInf(d, 0) && d > 0 {
			initPhoto()
			photo.SetExposureNumerator(1)
			photo.SetExposureDenominator(d)
		}
	}
}

That's one check for all the broken inputs (0, negative, NaN, subnormal) and no early return. Otherwise looks good to me.

[dschmidt]

I'm afraid my merge of #3198 caused a merge conflict here, can you rebase please? :)

[fredrikblau]

Rebased onto current main and resolved the conflict. The exposure-denominator guard and regression test are preserved. go test ./services/search/pkg/content and git diff --check pass locally.

[dschmidt]

LGTM, thanks! :)