feat(bernese): CODSPP-QC parse + a-priori/obs triage (RH-005 core)

[alfieprojectsdev]

RH-005 (core) — CODSPP-QC parse + a-priori/obs triage (readiness gap #9)

Stacked on #38 (branched off docs/bernese-training-notes); base retargets to main once #38 merges. Sanitizer-style detection core of the M-size RH-005; the auto-recovery actions are the remainder (below).

Problem

CODSPP (BPE jobs 301/302) is code single-point positioning run before the phase solution. Its per-cluster output (OUT/SPP_YYYYDDDs_NNN.OUT) reports, per station, an RMS OF UNIT WEIGHT and a coordinate correction table (NEW- A PRIORI). During the training week staff eyeballed those to separate two failure modes by hand:

• Bad a priori — high RMS and a large coordinate shift → the .CRD seed is wrong; CODSPP found the right spot. Free fix: re-seed from the CODSPP NEW coords and retry.
• Bad observations — high RMS but small shift → the data is bad; no auto-fix, alert a human.

Change — codspp_qc.py

• parse_codspp_output() → CodsppStation: station code, RMS OF UNIT WEIGHT, BAD/USED OBSERVATIONS, X/Y/Z NEW- A PRIORI deltas, and coord_shift_m (Euclidean magnitude). Missing fields degrade to None, never raise.
• classify_codspp() → ok / bad_apriori / bad_obs / unknown, with tunable rms_threshold_m (default 3) and shift_threshold_m (default 1). CODSPP code solutions normally sit ~1 m RMS — defaults must be tuned against a real network before unattended use.
• parse_codxtr_summary() reads the combined CODXTR worst-station MAX. RMS / MAX. BAD lines.

Verification

Parsed/classified against the real CUSV 0840 SPP block (RMS 1.03 m, shift 0.26 m → ok). test_codspp_qc.py +9, ruff + mypy clean. (Tests run via the already-synced rh-004 worktree venv + PYTHONPATH — deliberately no uv sync while the BPE final solve was running, after a uv sync I/O storm hung the run earlier today.)

RH-005 remainder (follow-up, tracked in backlog)

• Re-seed action (gap 🔒 Fix command injection in campaign_v6.py #9): on bad_apriori, rewrite the station's .CRD from CODSPP NEW coords and re-run — the applied/orchestrator layer.
• Tropo quarantine (gap 🔒 Fix potential DoS in TCP adapter by limiting buffer size #11): on PID 322 zenith-delay overflow, quarantine the malformed/short-baseline NON-IGS partner for the session and retry; never drop an IGS fiducial. Needs a failed-322 output sample to fix the parse against (none on disk this session).

Developed under the git-worktree protocol (.trees/rh-005-codspp-tropo).

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: c245111a-317a-411b-b0b3-d00db1207a7c

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feat/rh-005-codspp-tropo

---

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.}