Improve Rank robustness for malformed text/id rows

[hemanth-asirvatham]

Motivation

• Ranking failed when non-string or missing values were hashed (calling .encode() on floats/NA), causing crashes during identifier derivation and downstream indexing.
• Ensure ranking and related flows tolerate malformed rows by dropping or skipping them safely and informing users when rows are removed.
• Keep existing behavior for valid text while avoiding silent coercions that can produce subtle bugs.

Description

• Added helpers _is_missing_scalar and _hash_text_identifier to safely detect missing scalars and produce stable sha1 identifiers only for valid text-like inputs.
• Hardened seeding logic in _seed_ratings_from_rate to use safe hashing and drop invalid keys instead of raising.
• Modified both non-recursive (run) and recursive (_run_recursive) ranking entry paths to filter out malformed rows before processing, print a summary message when rows are dropped (count and percentage), and reset_index after filtering to avoid indexing issues.
• Added safe handling for the case where all rows are filtered out by returning and saving an appropriately shaped empty result (with attribute columns and _raw/_se columns present).
• Adjusted resume/checkpoint logic to use the safe hashing helper when comparing against previously saved identifiers.
• Added regression tests to tests/test_basic.py verifying malformed-row dropping behavior in standard and recursive modes.

Testing

• Ran targeted tests for the new behavior: pytest -q tests/test_basic.py -k "rank_drops_malformed_rows or recursive_rank_drops_malformed_rows or rank_outputs_standard_errors" which passed.
• Ran the broader rank-related subset: pytest -q tests/test_basic.py -k "rank and not screenshot" which passed (rank-related tests succeeded).
• All automated tests exercised during development for the rank changes passed after fixes; no failing automated tests remain for the modified areas.

---

Codex Task

[github-actions]

Thank you for your submission, we really appreciate it. Like many open-source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution. You can sign the CLA by just posting a Pull Request Comment same as the below format.

---

I have read the CLA Document and I hereby sign the CLA

---

_{You can retrigger this bot by commenting recheck in this Pull Request. Posted by the CLA Assistant Lite bot.}