Request type

Extension of existing module

Request description

This function allows users to link a metadata file and only download the newest (not already included) sequences/data. Done by comparing accession IDs. Also useful for when the API breaks and needs to be re-run from a certain checkpoint.

Feature: Baseline Metadata Deduplication with Merge Control

Problem Statements

Problem 1 (Pagination Failure Recovery): When API fetches fail mid-pagination, users lose data and must restart. Solution: Save partial metadata → user resumes with --baseline + --merge-results to combine new + old results.

Problem 2 (Incremental Updates): Users want to re-query without re-downloading data they already have. Solution: Provide --baseline with --no-merge to get only new sequences.

Proposed Solution

Single unified feature: baseline metadata deduplication

• Accepts baseline file (CSV/JSON/text) with accessions to skip
• User controls output: --merge-results (combined) or --no-merge (new only)
• On API failure: save partial metadata, suggest resume command in summary
• Works for both recovery AND incremental updates with same mechanism

Implementation Details

Data Flow

User Query + Optional --baseline file
    ↓
Fetch Metadata from API
    ↓
Deduplicate vs Baseline (skip accessions in file)
    ↓
Download Sequences for New Accessions Only
    ↓
Output: --merge-results → Combined | --no-merge → New Only

On API Failure

When pagination fails, the system:

1. Saves all metadata fetched so far to {virus}_partial_{timestamp}.csv
2. Saves this path in command summary
3. Suggests recovery command: gget virus {virus} --baseline {partial_file} --merge-results -o output/

Baseline File Format

Users provide metadata (CSV/JSON/text) with accession column:

accession,source,date
NC_045512.2,NCBI,2025-12-01
MN908947.3,local,2025-12-01

Function Signature

def virus(
    virus,
    baseline_metadata=None,  # Optional: accessions to skip
    merge_results=True,      # Default: merged output
    ...
):

Processing Flow

1. Fetch: Get all metadata from API (with retries on transient failures)
2. Deduplicate: Skip accessions already in baseline file
3. Download: Only new accessions
4. Merge:
   • --merge-results: Single output file with new + baseline combined
   • --no-merge: Separate files (new.csv, baseline_provided.csv) with clear labeling

Acceptance Criteria

• New parameter baseline_metadata accepts CSV/JSON/text files
• New parameter merge_results (default True) controls output format
• Baseline accessions correctly extracted and deduplicated
• On API failure: save partial metadata automatically
• Command summary shows recovery command with baseline file path
• Only new accessions downloaded (bandwidth savings)
• Output files properly labeled when --no-merge:
  • {virus}_new.csv (sequences from API)
  • {virus}_baseline_provided.csv (reference copy of baseline used)
  • {virus}_merged.csv (when using --merge-results)
• Comprehensive logging: accessions skipped, downloaded, merged
• Unit tests: baseline parsing (CSV/JSON/text), deduplication, merge logic

Related Code Sections

• virus() - lines 4668-5637
• fetch_virus_metadata() - lines 830-1320
• save_command_summary() - where recovery suggestion added
• download_sequences_by_accessions() - skips baseline accessions

Benefits

✅ Single Solution, Two Problems - Same feature handles both failures and incremental updates
✅ Simple Recovery - One suggested command in error summary
✅ User Control - Merge option for flexible output
✅ Bandwidth Efficient - Only download new sequences
✅ Transparent - Clear logging of deduplication
✅ Minimal Code - No special checkpoint logic, just baseline deduplication

Potential Concerns & Mitigations

Usage Examples

Scenario 1: API Failure → Recovery

# Run 1: Starts fetching, fails at page 28 of 35
$ gget virus SARS-CoV-2 -o output/
ERROR: Connection lost at page 28
Partial metadata saved: output/SARS-CoV-2_partial_20260205_120530.csv
Recovery command: gget virus SARS-CoV-2 --baseline output/SARS-CoV-2_partial_20260205_120530.csv --merge-results -o output/

# Run 2: User runs suggested command
$ gget virus SARS-CoV-2 --baseline output/SARS-CoV-2_partial_20260205_120530.csv --merge-results -o output/
Baseline loaded: 28,000 accessions from partial fetch
New accessions from API: 22,000
✅ output/SARS-CoV-2_virus_sequences.csv (merged: 50,000 total)

Scenario 2: Incremental Update (Merged)

$ gget virus Zika --baseline previous_zika.csv --merge-results -o output/
Baseline loaded: 2,000 accessions
API returned: 2,500 records
New accessions: 500 (downloaded)
✅ output/zika_virus_sequences.csv (merged: 2,500 total)

Scenario 3: Incremental Update (New Only)

$ gget virus Zika --baseline previous_zika.csv --no-merge -o output/
Baseline loaded: 2,000 accessions
API returned: 2,500 records
New accessions: 500 (downloaded)
✅ output/zika_virus_new.csv (500 sequences)
✅ output/zika_virus_baseline_provided.csv (reference)

Example command

`gget virus {virus} --baseline {partial_file} --merge-results -o output/`

Example return value

No response