fix: score a tuple of pairs in rerank_pairs instead of treating it as one pair

[michaelkillgta]

Fixes #796

OnnxCrossEncoderModel._rerank_pairs treated every tuple input as a single (query, document) pair and wrapped it in a list. A tuple of pairs (still a valid Iterable[tuple[str, str]]) became one malformed pair, and the tokenizer raised TypeError: TextInputSequence must be str.

Change

• Treat a tuple as a single pair only when it is exactly two strings. Any other tuple is converted to a list of pairs, so it also gets the same small-input fast path that lists already get.
• Behaviour for lists, generators and a single (query, document) tuple is unchanged.

Test

• test_rerank_pairs_tuple_input checks that a tuple of two pairs and a tuple of three pairs give the same scores as the equivalent list, and that a single pair still works. It fails on main with the TypeError above and passes with this change.

All Submissions:

• Have you followed the guidelines in our Contributing document?
• Have you checked to ensure there aren't other open Pull Requests for the same update/change?

New Feature Submissions:

• Does your submission pass the existing tests?
• Have you added tests for your feature?
• Have you installed pre-commit ...? (Not installed as a hook. I ran the pinned ruff 0.3.4 check and format directly.)

Local runs: tests/test_text_cross_encoder.py and tests/test_custom_models.py give 15 passed. The one failure, test_rerank, already fails on main on my machine because of the bge-reranker-v2-m3-int8 canonical scores, so it is unrelated. ruff 0.3.4 (the pre-commit pin) check and format pass, and mypy fastembed with the CI flags passes.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

• Configuration used: Organization UI
• Review profile: CHILL
• Plan: Advanced
• Run ID: c71b3413-aee2-416f-9b90-367b367516d6

---

📥 Commits

Reviewing files that changed from the base of the PR and between d076f08 and 04e9688.

---

📒 Files selected for processing (2)

• fastembed/rerank/cross_encoder/onnx_text_model.py
• tests/test_text_cross_encoder.py

---

Included review availability: This review used your included allowance. Your plan provides up to 8 included reviews per hour; 7 remain after this review.

---

---

📝 Walkthrough

Walkthrough

_rerank_pairs now treats a tuple of exactly two strings as one query-document pair. It converts other tuples to lists of pairs. Tests check that tuples of two and three pairs produce the same scores as equivalent lists, and that a single query-document tuple remains accepted.

Priority: ➖ Normal

Estimated code review effort: 2 (Simple) | ~10 minutes

Change: Bug fix · Severity of issue fixed: Medium

---

Merge Risk: ⚪ Minimal · up to 04e96

Tuples of query-document pairs are now scored like equivalent lists, and a single (query, document) tuple still works. No merge-blocking risk was identified.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage		Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 4 functions across 2 files.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Linked Issues check		Issue #796 requires rerank_pairs() to score a tuple of (query, document) pairs like an equivalent list and to preserve a single pair tuple. The change in OnnxCrossEncoderModel._rerank_pairs trea…
Out of Scope Changes check		The diff contains only the tuple-input fix in fastembed/rerank/cross_encoder/onnx_text_model.py and its regression test in tests/test_text_cross_encoder.py. Both changes directly support issue #79…
Title check		The title clearly and concisely identifies the main change: correcting tuple-of-pairs handling in rerank_pairs.
Description check		The description directly explains the tuple-handling bug, the implementation change, preserved behavior, tests, and validation results.

---

---

• Fix all pre-merge checks with AI

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create a new PR

---

---

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

---

_{Comment @coderabbitai help to get the list of available commands.}