1. Initialize ChebiSDF¶

On first initialization, an index will be built (takes ~15 seconds). Subsequent loads use the cached index.

```{code-cell} ipython3 from provesid.chebi import ChebiSDF

Initialize (will use cached index if available)¶

chebi_sdf = ChebiSDF()

Get database statistics¶

stats = chebi_sdf.get_database_stats() print("ChEBI SDF Database Statistics:") print("=" * 50) for key, value in stats.items(): print(f"{key:30s}: {value:>15,}")

## 2. Query by ChEBI ID

```{code-cell} ipython3
# Get water (CHEBI:15377)
water = chebi_sdf.get_compound_by_id("CHEBI:15377")

print(f"ChEBI ID: {water['ChEBI ID']}")
print(f"Name: {water['ChEBI NAME']}")
print(f"Formula: {water.get('FORMULA', 'N/A')}")
print(f"Mass: {water.get('MASS', 'N/A')}")
print(f"SMILES: {water.get('SMILES', 'N/A')}")
print(f"InChI: {water.get('INCHI', 'N/A')}")
print(f"InChIKey: {water.get('INCHIKEY', 'N/A')}")
print(f"Star Rating: {water.get('STAR', 'N/A')}")

3. Search by Name¶

```{code-cell} ipython3

Exact name search¶

results = chebi_sdf.search_by_name("glucose", exact=True) print(f"Found {len(results)} compound(s) with exact name 'glucose'") for result in results[:3]: print(f" - {result['ChEBI ID']}: {result['ChEBI NAME']}")

print()

Partial name search¶

results = chebi_sdf.search_by_name("glucose", exact=False) print(f"Found {len(results)} compound(s) with 'glucose' in the name") for result in results[:5]: print(f" - {result['ChEBI ID']}: {result['ChEBI NAME']}")

## 4. Search by CAS Number

```{code-cell} ipython3
# Search for aspirin (CAS: 50-78-2)
results = chebi_sdf.search_by_cas("50-78-2")

if results:
    aspirin = results[0]
    print(f"ChEBI ID: {aspirin['ChEBI ID']}")
    print(f"Name: {aspirin['ChEBI NAME']}")
    print(f"Formula: {aspirin.get('FORMULA', 'N/A')}")
    print(f"\nSynonyms:")
    if 'SYNONYM' in aspirin:
        synonyms = aspirin['SYNONYM'].split(';')
        for syn in synonyms[:5]:
            print(f"  - {syn.strip()}")

5. Search by Structure Identifiers¶

```{code-cell} ipython3

Search by InChIKey (water)¶

inchikey = "XLYOFNOQVPJJNP-UHFFFAOYSA-N" result = chebi_sdf.search_by_inchikey(inchikey)

if result: print(f"Found compound: {result['ChEBI NAME']} ({result['ChEBI ID']})")

Search by molecular formula¶

results = chebi_sdf.search_by_formula("H2O") print(f"\nFound {len(results)} compound(s) with formula H2O") for r in results[:3]: print(f" - {r['ChEBI ID']}: {r['ChEBI NAME']}")

## 6. Search by Synonym

```{code-cell} ipython3
# Search for compounds with "acetylsalicylic acid" as a synonym
results = chebi_sdf.search_by_synonym("acetylsalicylic acid", exact=False)

print(f"Found {len(results)} compound(s)")
for result in results:
    print(f"\nChEBI ID: {result['ChEBI ID']}")
    print(f"Name: {result['ChEBI NAME']}")
    print(f"Formula: {result.get('FORMULA', 'N/A')}")

7. Batch Operations and DataFrame Export¶

```{code-cell} ipython3

Get multiple compounds at once¶

chebi_ids = ["CHEBI:15377", "CHEBI:16236", "CHEBI:17234"] # water, ethanol, glucose

Export to DataFrame¶

df = chebi_sdf.export_to_dataframe(chebi_ids)

print("DataFrame with selected compounds:") print(df[['ChEBI ID', 'ChEBI NAME', 'FORMULA', 'MASS', 'STAR']])

## 8. Working with High-Quality Compounds

ChEBI assigns star ratings (1-3) to compounds, where 3 stars indicates the highest quality/completeness.

```{code-cell} ipython3
# Note: This will take a few minutes as it scans all compounds
# Uncomment to run:

# three_star_ids = chebi_sdf.filter_by_star_rating(min_stars=3)
# print(f"Found {len(three_star_ids)} compounds with 3-star rating")

# # Export first 100 to DataFrame
# df_high_quality = chebi_sdf.export_to_dataframe(three_star_ids[:100])
# print(df_high_quality.head())

9. Accessing External Database Links¶

```{code-cell} ipython3

Get water and show all available database links¶

water = chebi_sdf.get_compound_by_id("CHEBI:15377")

print("External Database Links for Water:") print("=" * 50) for key, value in water.items(): if 'Database Links' in key or 'Registry Number' in key: print(f"{key}: {value}") ```

Summary¶

The ChebiSDF class provides:

Fast queries using pre-built index (~190,000 compounds in <1 second after indexing)
Multiple search methods: by ID, name, synonym, CAS, InChI, InChIKey, formula
Batch operations: get multiple compounds at once
DataFrame export: easy integration with pandas for analysis
Offline access: no internet required after SDF file is downloaded
Complete data: structure, properties, synonyms, and 80+ database cross-references