Exporting Results

Export your batch processing results in multiple formats for analysis, reporting, and data exchange. ChemAudit supports CSV, Excel, SDF, JSON, and PDF exports with flexible filtering options.

Available Export Formats

Format	Extension	Use Case	Features
CSV	`.csv`	Spreadsheet analysis, data pipelines	Simple tabular data
Excel	`.xlsx`	Formatted reports, presentations	Conditional coloring, summary sheet
SDF	`.sdf`	Structure-data exchange	Compatible with chemistry tools
JSON	`.json`	Programmatic processing	Full data fidelity, nested structures
PDF	`.pdf`	Professional reports, archival	Charts, molecule images, statistics
Fingerprint Matrix	`.zip`	ML pipeline input	Morgan/MACCS/RDKit FPs in CSV + numpy
Deduplicated Set	`.zip`	Dataset curation	Summary + annotated CSVs with group IDs
Scaffold-Grouped	`.csv`	SAR analysis	Murcko scaffolds + group assignments
Property Matrix	`.zip`	Comprehensive analysis	Flat CSV + 4-sheet Excel

How to Export

Web Interface

Complete a batch processing job
Click the Export button
Select export format
(Optional) Apply filters:
- Minimum/maximum validation score
- Success or error status
- Specific molecule indices
Click Download

The file downloads automatically with a descriptive filename including job ID and timestamp.

API - Export All Results

# Export all results as CSV
curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=csv" -o results.csv

# Export as Excel
curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=excel" -o results.xlsx

# Export as SDF
curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=sdf" -o results.sdf

# Export as JSON
curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=json" -o results.json

# Export as PDF report
curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=pdf" -o report.pdf

API - Export with Filters

# Export molecules with validation score >= 80
curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=csv&score_min=80" -o high_quality.csv

# Export only successful validations
curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=excel&status=success" -o successful.xlsx

# Export specific score range
curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=sdf&score_min=70&score_max=90" -o medium_scores.sdf

API - Export Selected Molecules

For small selections (up to 200 indices), use GET with comma-separated indices:

curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=csv&indices=0,1,5,23,42" -o selected.csv

For large selections (> 200 indices), use POST with JSON body:

curl -X POST "http://localhost:8001/api/v1/batch/{job_id}/export?format=json" \
  -H "Content-Type: application/json" \
  -d '{"indices": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]}' \
  -o selected.json

Format Details

CSV Export

Simple tabular format with these columns:

Index
SMILES
Name (if provided)
Validation score
Status (success/error)
Number of issues
ML-readiness score
QED score
Safety pass/fail
Alert counts (PAINS, BRENK, etc.)
Molecular formula
Molecular weight

Pros:

Universal compatibility
Easy to import into spreadsheets or databases
Small file size

Cons:

No nested data structures
No formatting or charts
Limited molecule information

Excel Export

Formatted spreadsheet with two sheets:

Results Sheet:

All molecule data in table format
Conditional formatting (red for errors, green for high scores)
Frozen header row
Auto-sized columns

Summary Sheet:

Job statistics
Score distribution
Alert summary
Processing metadata

Pros:

Professional appearance
Ready for presentations
Includes summary statistics
Conditional coloring highlights issues

Cons:

Larger file size than CSV
Requires Excel or compatible software

SDF Export

Structure-Data File format with molecular structures and properties:

Molecule_0
  RDKit          2D

 13 13  0  0  0  0  0  0  0  0999 V2000
    ... coordinates ...
M  END
> <validation_score>
95

> <qed>
0.65

> <ml_readiness_score>
85

$$$$

Pros:

Preserves 2D/3D structure coordinates
Compatible with chemistry software (RDKit, MOE, Schrodinger)
Can include custom properties

Cons:

Larger file size
Not human-readable for data analysis

Chemistry Tools

SDF is the best format for importing into molecular modeling, docking, or cheminformatics tools.

JSON Export

Complete data in JSON format with nested structures:

[
  {
    "index": 0,
    "smiles": "CCO",
    "name": "Ethanol",
    "status": "success",
    "validation": {
      "overall_score": 95,
      "issues": [],
      "all_checks": [...]
    },
    "alerts": {
      "total_alerts": 0,
      "alerts": []
    },
    "scoring": {
      "ml_readiness": {...},
      "druglikeness": {...},
      "admet": {...}
    }
  }
]

Pros:

Complete data preservation
Nested structures maintained
Easy to parse programmatically
Standard format for APIs

Cons:

Larger file size
Not suitable for spreadsheet analysis

PDF Report

Professional report document with:

Executive Summary
- Job metadata (date, molecule count, processing time)
- Overall statistics
- Key findings
Score Distribution Charts
- Validation score histogram
- ML-readiness distribution
- QED distribution
Molecule Gallery
- Structure images (up to 100 molecules)
- Key properties
- Validation scores
- Alert indicators
Failed Molecules Section
- Error messages
- Failure reasons
- Recommendations
Alert Summary
- Alert counts by catalog
- Most common patterns
- Risk assessment
Processing Metadata
- Configuration used
- Processing time
- Worker statistics

Pros:

Professional presentation
Self-contained (no external dependencies)
Suitable for archival and sharing
Includes visualizations

Cons:

Largest file size
Not suitable for further processing
Limited to summary information for large batches

PDF Size Limits

For batches larger than 1,000 molecules, PDF exports include only summary statistics and the first 100 molecules to keep file size manageable.

Fingerprint Matrix Export

Generates a ZIP containing 9 files — 3 fingerprint types in 3 formats each:

Fingerprint	Bits	File Prefix
Morgan ECFP4 (radius 2)	2,048	`morgan`
MACCS keys	167	`maccs`
RDKit path-based	2,048	`rdkit`

Each fingerprint is exported as .csv, .npy (numpy array), and .npz (compressed numpy). Molecules that failed parsing are skipped. Ideal for feeding directly into ML pipelines.

curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=fingerprint" -o fingerprints.zip

Deduplicated Set Export

Generates a ZIP with two CSV files:

dedup_summary.csv — One row per unique canonical SMILES group: group_id, representative SMILES, group size, duplicate indices, dedup level
dedup_annotated.csv — All molecules with group_id, is_representative flag, dedup_level, validation score, and InChIKey

Ideal for dataset curation — quickly identify and remove duplicates while preserving provenance.

curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=dedup" -o deduplicated.zip

Scaffold-Grouped Export

Single CSV with Murcko scaffold assignments:

scaffold_smiles — the Murcko scaffold SMILES (empty string for acyclic molecules)
scaffold_group — integer group ID (0 for acyclic, 1+ for scaffold groups)
Plus: index, SMILES, name, status, overall score, ML-readiness score, InChIKey

Ideal for SAR analysis and scaffold-based dataset organization.

curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=scaffold" -o scaffolds.csv

Property Matrix Export

Generates a ZIP with two files:

properties_flat.csv — All properties as columns in a single flat table
properties.xlsx — Multi-sheet Excel workbook:
- Descriptors: MW, LogP, TPSA, HBD, HBA, Rotatable Bonds, Aromatic Rings, Fsp3, Heavy Atom Count, Ring Count
- Scores: Validation, ML-readiness, QED, SA, NP-likeness, Lipinski violations
- Alerts: Per-catalog alert counts (PAINS, BRENK, NIH, Glaxo) and alert names
- Identifiers: SMILES, name, InChIKey, status

curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=property_matrix" -o properties.zip

PDF Section Selection

PDF exports support selecting specific sections via the sections parameter:

Section Key	Content
`validation_summary`	Overall statistics and pass rates
`score_distribution`	Score histogram charts
`alert_frequency`	Alert count bar charts
`chemical_space`	Chemical space scatter plot
`scaffold_treemap`	Scaffold frequency visualization
`statistics`	Property statistics tables
`correlation_matrix`	Property correlation heatmap
`mmp_pairs`	Matched molecular pair summary

All sections are included by default. To select specific sections:

curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=pdf&sections=validation_summary,score_distribution" -o summary.pdf

Excel Structure Images

Add include_images=true to embed 2D structure PNG images in the Excel export:

curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=excel&include_images=true" -o results_with_images.xlsx

Score Filters

Parameter	Type	Description
`score_min`	integer (0-100)	Minimum validation score
`score_max`	integer (0-100)	Maximum validation score

Example:

# Export only high-quality molecules (score >= 90)
curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=csv&score_min=90"

Status Filters

Parameter	Values	Description
`status`	`success`, `error`, `warning`	Processing outcome

Example:

# Export only failed molecules for debugging
curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=csv&status=error"

Molecule Selection

Parameter	Type	Description
`indices`	string or array	Specific molecule indices

GET Example (small selections):

curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=sdf&indices=0,1,2,3"

POST Example (large selections):

curl -X POST "http://localhost:8001/api/v1/batch/{job_id}/export?format=json" \
  -d '{"indices": [0,1,2,3,4,5,6,7,8,9]}'

Combining Filters

Filters can be combined for precise exports:

# High-quality successful molecules with specific indices
curl "http://localhost:8001/api/v1/batch/{job_id}/export?format=excel&status=success&score_min=80&indices=0,5,10,15"

Best Practices

Choose the right format:
- CSV for data analysis
- Excel for reports and presentations
- SDF for chemistry software
- JSON for programmatic processing
- PDF for archival and sharing
Use filters to reduce file size:
- Export only relevant molecules
- Filter by score to focus on quality
- Separate successes and errors
Export multiple times for different purposes:
- CSV for your database
- Excel for stakeholder reports
- SDF for molecular modeling
- PDF for documentation
Preserve raw data:
- Keep original batch results for reproducibility
- Export JSON for complete data preservation

Next Steps

Batch Processing — Process large datasets
Batch Analytics — Analytics that power advanced exports
Subset Actions — Export selected molecules only
API Reference — Full export API documentation

Available Export Formats​

How to Export​

Web Interface​

API - Export All Results​

API - Export with Filters​

API - Export Selected Molecules​

Format Details​

CSV Export​

Excel Export​

SDF Export​

JSON Export​

PDF Report​

Fingerprint Matrix Export​

Deduplicated Set Export​

Scaffold-Grouped Export​

Property Matrix Export​

PDF Section Selection​

Excel Structure Images​

Filter Options​

Score Filters​

Status Filters​

Molecule Selection​

Combining Filters​

Best Practices​

Next Steps​

Available Export Formats

How to Export

Web Interface

API - Export All Results

API - Export with Filters

API - Export Selected Molecules

Format Details

CSV Export

Excel Export

SDF Export

JSON Export

PDF Report

Fingerprint Matrix Export

Deduplicated Set Export

Scaffold-Grouped Export

Property Matrix Export

PDF Section Selection

Excel Structure Images

Filter Options

Score Filters

Status Filters

Molecule Selection

Combining Filters

Best Practices

Next Steps