[t0mdavid-m]

Needs openms-insight with t0mdavid-m/OpenMS-Insight#32 released. That PR adds the table options this one uses (search, search_placeholder, numeric_filters). requirements.txt still pins openms-insight==0.4.0. After the release, this PR needs one more commit that bumps the pin. Until then, tests/test_sequence_display.py::test_filtered_table_adds_q_values_and_a_filter_menu fails against 0.4.0. Everything else passes either way.

Before: the PSM and Identified proteins pages only had the Insight Table's per-column filter dialog:

• You couldn't search the whole table.
• You couldn't tighten the FDR, drop one-hit proteins, require a protein in several replicates of a condition, or hide contaminants on the protein table.
• The PSM table showed only the score, which is Percolator's PEP, while the page's info box called it the q-value.

After: each table filters itself, from its own toolbar and filter menu. There is no separate filter bar above it.

• PSM table:
  • The search box in the toolbar searches the sequence in any modification form (M(ox), M[Oxidation] and M(Oxidation) all match) and the protein accessions.
  • The filter button opens the dialog on q-value thresholds. "1 % (run)", the run's own FDR, comes first, then 0.5 / 0.1 / 0.01 %, so a threshold can only tighten the run's.
  • A new q-value column sits next to Score. m/z and RT are rounded for display, and the info box says what Score and q-value are.
• Protein table:
  • The search box searches accession, gene and description.
  • The dialog offers peptides ≥ 2 / 3 / 5 (distinct peptides) and replicates ≥ 1 … n. "Replicates" is the most files of one condition with a PSM of the protein, using the conditions from Describe experiment, and is shown in a new replicates column. A fractionated run counts samples, not fractions.
  • The key figures above the table count what its filters leave, without contaminants.
• Contaminants keep openms-insight's toggle_filters checkbox in each table's title row. It is on by default for proteins and off for PSMs.
• Every filter in force shows as a chip under the toolbar. Each chip can be removed, and "Clear all" removes them all. The filters stay when you leave the page and come back.

Dropped: the protein page's PSM q-value threshold. It filtered the PSMs underneath the protein table, so it changed each protein's peptide and PSM counts, its coverage and the linked peptide and PSM tables. A filter on a column of the protein table can't express that. Faking it with a per-protein "best q-value" column would keep proteins whose counts still include PSMs above the threshold, which is misleading. The PSM table keeps its q-value threshold.

How: the search and the thresholds are openms-insight Table options, applied in Python over every row through the table's pagination state. Neither a search nor a threshold builds a new cache.

• PSM page: sequence_display.filtered_table_id() builds one table variant per run and sequence form from the workflow's cached PSM table, as the modification-display variants already do. It adds q_value (read from the idXML in id_idx order) and not_contaminant, and declares search=[sequence, sequence_display, proforma, protein_accession], numeric_filters={"q_value": ...} and toggle_filters. The thresholds offered come from result_filters.q_suggestions().
• Protein page: it passes the whole protein table to show_insight_table with search, numeric_filters (protein_numeric_filters()) and toggle_filters. show_insight_table now returns the Table, so the key figures come from Table.export_data(state), the rows the viewer's filters leave.
• src/common/filter_bar.py is removed. The no-Streamlit logic stays in src/common/result_filters.py.

Proteins: dialog with thresholds	Proteins: ≥ 2 peptides, ≥ 2 replicates	Proteins: plus search

PSMs: q-value thresholds	PSMs: search + q ≤ 0.001	PSMs: q ≤ 0.0001 leaves none

Video (mp4): https://raw.githubusercontent.com/OpenMS/quantms-web/claude/media-nav/r6-table-menu.mp4

Testing

• Real LFQ run on 4 demo files (2 per condition, UPS2_ratio01 / UPS2_ratio10, yeast + UPS2 database, 1 % PSM FDR), in headless Chromium with the openms-insight branch on PYTHONPATH.
• Identified proteins (key figures: proteins / peptides / PSMs):
  • no filters: 627 / 850 / 3,979;
  • ≥ 2 peptides and ≥ 2 replicates: 166 / 404 / 2,577, the same as the previous filter bar gave;
  • plus search "dehydrogenase": 11 / 25 / 239;
  • "Clear all": 627 / 850 / 3,979.
• PSM table, file B000248: 1,013 PSMs; "CAH2" leaves 23. Every PSM that passed in this run has q = 0.00095, so q ≤ 0.001 keeps all 23 and q ≤ 0.0001 leaves none ("Showing 0 rows"). Removing that chip brings back the 23.
• tests/test_result_filters.py covers:
  • q thresholds start at the run and never loosen it, including a run without a PSM filter;
  • the protein thresholds offered;
  • replicates per condition, and per sample for fractions;
  • the protein flags and the column placement.
• test_filtered_table_adds_q_values_and_a_filter_menu checks:
  • q-values join by id_idx;
  • the variant's search / threshold / toggle config;
  • the variant is built once;
  • the table's own ProForma search and q threshold.
• python -m pytest tests/ test_gui.py with the openms-insight branch: 363 passed, 2 skipped. With the pinned 0.4.0: the one test named above fails, all others pass. python -m py_compile content/*.py src/common/*.py: clean. pylint --errors-only: only the existing E0611 errors on pyopenms imports.

Note: with 4 or more files the protein table folds its per-file columns into the rows (Tabulator's responsive collapse). That happens on main too, and this PR doesn't change it.

🤖 Generated with Claude Code

https://claude.ai/code/session_01D5Wo2FHPpi5a8X6GppNxEk

[coderabbitai]

Important

• Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

Run configuration

• Source mode: Auto
• Repository: Unavailable
• Branch: Unavailable
• File: No configuration file from this repository applied
• Commit: Unavailable
• Layers: Organization UI
• Review profile: CHILL
• Plan: Advanced
• Run ID: 290001db-aa85-4ae1-a1f4-a33eaf027e61

• 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

[t0mdavid-m]

Three checks are red on 1aaefc8, and none of them needs a code change in this PR:

• test (ubuntu-latest, 3.11): one test fails, tests/test_sequence_display.py::test_filtered_table_adds_q_values_and_a_filter_menu, with KeyError: 'numeric_filters'; the other 369 pass. This is expected until openms-insight is released with Table: search box, quick numeric thresholds and removable filter chips t0mdavid-m/OpenMS-Insight#32. CI installs the pinned openms-insight==0.4.0, which doesn't have the search / numeric_filters Table options. Once that release is out, I'll push the pin bump here. The test passes locally with the fix(results): support pyOpenMS 3.5.0 IdXMLFile.load signature #32 branch.
• Desktop Linux-x86_64: this failure isn't from this PR. The electron-rescoring.js scenario stops at "Start Workflow", the same failure as on main since Rescore with predicted spectra and retention times (opt-in) #83. A workflow run with predicted-spectrum rescoring doesn't finish in the desktop app; that is being fixed in Run local workflows in a fresh interpreter; desktop e2e waits for the whole run #87. All 24 pages opened, and electron-workflows.js passed, including the PSM and protein tables. Once the fix is in main, I'll merge main into this branch.
• test-nginx (simple, arm64, ubuntu-24.04-arm): this is the kind cluster, not the app. The ingress-nginx controller was still ContainerCreating when the job gave up, and no app resources were in the openms namespace yet. This PR changes no k8s or Docker files. I'll re-run the job once when the run finishes; it is still running now.

---

Generated by Claude Code

[t0mdavid-m]

fix ci