Results

The Results page is the detail view for a single BLAST search. It opens when you click a row on BLAST Jobs or follow a /blast/jobs/<jobId> link from a submit toast.

Overview

The page is built around three regions:

1. Job header — NCBI-style metadata about the search.
2. Sticky tab bar — switches the body between Descriptions, Graphic Summary, Alignments, Taxonomy, Files, Comparison, and Run details.
3. Body — the table, panel, or timeline for the active tab.

Job Header

The header is laid out in four lines, top to bottom:

1. A < BLAST Jobs back link plus a Help menu with links to NCBI BLAST documentation.
2. Job title with Cancel, Edit search, and Duplicate actions on the right. Cancel is shown only while the search is still running.
3. Search ID: … and Created: … with Copy ID and a Download all results ▾ combo (BLAST XML, JSON, TSV, ZIP archive).
4. Program: blastn Database: … Query: … Molecule: dna Query length: … — the four facts a researcher checks first to confirm "yes, this is my search".

Cluster, region, and storage account live on the Run details tab, not in the header.

The Reproducibility action downloads a portable JSON package containing the canonical submit snapshot, provenance, Methods citations, workflow-manager modules, and the result manifest when one is available. It intentionally omits the raw FASTA and execution idempotency/correlation identifiers. Use the separate Edit search action when you need to reload the original query into the submit form.

Tabs

Tab	What it shows
Descriptions	NCBI-style "Sequences producing significant alignments" table — hit accession, definition, length, score, E-value, identity, coverage. Sort and per-hit links to NCBI.
Graphic Summary	The familiar coloured-bar plot showing where each hit aligns along the query.
Alignments	Pairwise alignment view for each hit, with score/E-value/identity per HSP.
Taxonomy	Organism rollup of the current hit set, grouped by lineage.
Files	The raw result files in Storage — merged XML, per-shard outputs, support files, and (for failed runs) debug logs. Click Download to stream a file through the API.
Comparison	Compare this completed result set with another completed search that used the same program and database. Shows added, removed, changed, and unchanged query/subject hits.
Run details	Execution timeline, per-pod / per-shard status, sharding mode, warmup state, and the resolved BLAST command.

The active tab is encoded in ?tab=… so deep links survive a reload and browser back/forward work as expected.

Comparison reads both jobs' existing result files through the API's normal bounded Storage path. A warning appears when a byte, file, hit, or response-item limit makes the comparison partial; the UI never presents that subset as a full result-set equality claim.

For a split-query parent, Run details includes a Shard details table with each child job's group, query filename, status, duration, effective search space, and sanitized failure text. It polls while any child is active and stops automatically after every child reaches a terminal state. Non-split jobs do not make the shard-detail request or render the table.

The table returns at most 1,000 child rows. A visible warning marks larger jobs as truncated and clarifies that the displayed summary covers only those rows. A completed-job Run details deep link survives refresh; the page switches to Descriptions automatically only when it observes an active job complete.

Run States You Will See

Body state	When it appears	What to do
Result tables (Descriptions / Graphic / Alignments / Taxonomy)	The merged result file is published and parsable.	Read the hits; download from the header combo when needed.
Search failed during the <step> step.	The job ended in a failed state.	Open Run details for diagnostics; use the terminal only if you need to inspect logs directly.
Running placeholder	The job is still executing and no result file has been published yet.	Stay on the page — it polls automatically and switches to the result view when artefacts appear.
Cannot load results — missing Azure configuration.	No subscription or storage account is selected.	Open the Dashboard and finish the setup wizard.
Storage locked panel	publicNetworkAccess is Disabled and the browser is outside the platform VNet (typical for a developer laptop).	This is the expected production posture. Either run the operation from inside the deployed Container App or use the local-debug helper documented in Storage Network Isolation & Browser ↔ Storage Proxy.
No result files	The run completed without publishing any usable artefacts.	Check Run details for shard-level errors; consider re-submitting with a smaller query.

Failed runs still expose Edit search and Duplicate in the header so you can re-submit with a tweaked configuration.

Downloading Results

The Download all results combo in the header streams a packaged artefact through the API. The combo offers BLAST XML, JSON, TSV (descriptions), and a ZIP that bundles every shard output.

For individual files, switch to the Files tab and download row-by-row. Downloads always go through the API sidecar — the browser never receives a Storage SAS URL.

Screenshot Targets

Screenshots for this page are defined by this manifest target:

• results-desktop

Use a completed demo job with a small query (for example the monkeypox sample used in the API Reference) so the Descriptions tab is populated. Mask UPNs, subscription IDs, full storage account names, and any private query identifiers before publishing.