API Documentation

run_full_pipeline()

run_full_pipeline(
    disease_id: str,
    entity_file: Optional[str] = None,
    knowledge_graph: Optional[str] = None,
    entity_embeddings: Optional[str] = None,
    relation_embeddings: Optional[str] = None,
    output_dir: Optional[str] = None,
    model: str = 'TransE_l2',
    top_n_diseases: int = 100,
    gamma: float = 12.0,
    threshold: float = 0.5,
    relation_type: str = 'GNBR::T::Compound:Disease',
    top_n_drugs: int = 1000
) -> None

Description: One-click complete analysis pipeline that generates annotated_drugs.xlsx with all analysis results.

Parameters:

• disease_id: Disease identifier (e.g., "MONDO:0004979")
• entity_file: Path to entity annotations file (optional, automatically downloaded from wandb if not specified)
• knowledge_graph: Path to knowledge graph file (optional, automatically downloaded from wandb if not specified)
• entity_embeddings: Path to entity embeddings file (optional, automatically downloaded from wandb if not specified)
• relation_embeddings: Path to relation embeddings file (optional, automatically downloaded from wandb if not specified)
• output_dir: Output directory for results
• model: KGE model type (default: 'TransE_l2')
• top_n_diseases: Number of similar diseases to consider (default: 100)
• gamma: Margin parameter for KGE training (default: 12.0)
• threshold: Drug filtering threshold (default: 0.5)
• relation_type: Relation type for drug-disease associations (default: 'GNBR::T::Compound:Disease')
• top_n_drugs: Number of drugs to analyze (default: 1000)

Returns:

None

Example:

core = DrugDiseaseCore()
core.run_full_pipeline(
    disease_id="MONDO:0004979",
    output_dir="results/",
    model='TransE_l2',
    top_n_diseases=50,
    gamma=12.0,
    threshold=0.5,
    top_n_drugs=100
)

predict_drugs()

predict_drugs(
    disease_id: str,
    entity_file: str,
    knowledge_graph: str,
    entity_embeddings: str,
    relation_embeddings: str,
    model: str,
    top_n_diseases: int,
    gamma: float,
    threshold: float,
    relation_type: str,
    output_file: str
) -> None

Description: Generate potential drug list using KGE models.

Parameters:

• disease_id: Target disease identifier
• entity_file: Path to entity annotations file
• knowledge_graph: Path to knowledge graph file
• entity_embeddings: Path to entity embeddings file
• relation_embeddings: Path to relation embeddings file
• model: KGE model type
• top_n_diseases: Number of similar diseases
• gamma: Margin parameter
• threshold: Prediction threshold
• relation_type: Relation type for drug-disease associations
• output_file: Output Excel file path

get_disease_name()

get_disease_name(disease_id: str, entity_file: str) -> str

Description: Get disease name from disease ID.

Parameters:

• disease_id: Disease identifier
• entity_file: Path to entity annotations file

Returns:

Disease name as string

get_drug_names()

get_drug_names(drug_ids: List[str], entity_file: str) -> List[str]

Description: Get drug names from drug IDs.

Parameters:

• drug_ids: List of drug identifiers
• entity_file: Path to entity annotations file

Returns:

List of drug names

DrugFilter()

filter_drugs()

filter_drugs(
    input_file: str,
    expression: str,
    output_file: str,
    sheet_names: Tuple[str, str] = ("annotated_drugs", "filtered_drugs")
) -> None

Description: Filter drugs based on logical expressions.

Parameters:

• input_file: Input Excel file path
• expression: Filter expression (e.g., "score > 0.6 and existing == False")
• output_file: Output Excel file path
• sheet_names: Tuple of (input_sheet, output_sheet) names

Example:

filter = DrugFilter()
filter.filter_drugs(
    input_file="results/annotated_drugs.xlsx",
    expression="score > 0.7 and num_of_shared_genes_in_path >= 1",
    output_file="results/filtered_drugs.xlsx"
)

Supported Filter Expressions:

• Numerical comparisons: >, <, >=, <=, ==, !=
• Logical operators: and, or, not
• Boolean fields: existing, is_key_gene
• String matching and pattern matching

Visualizer(disease_id: str, disease_name: str, embed_images: bool = True)

create_visualization()

create_visualization(
    data_file: str,
    viz_type: str,
    output_file: str,
    sheet_names: Tuple[str, str] = ("annotated_drugs", "filtered_drugs")
) -> str

Description: Generate specific visualization chart.

Parameters:

• data_file: Input data file path
• viz_type: Visualization type (see supported types below)
• output_file: Output file path
• sheet_names: Tuple of sheet names for input data

Supported Visualization Types:

• score_distribution - Score distribution histogram
• score_boxplot - Score distribution by existing status
• disease_similarity - Disease similarity heatmap
• network_centrality - Network centrality analysis
• shared_genes_pathways - Gene and pathway overlap
• drug_similarity_network - Drug similarity network
• shared_gene_count - Shared gene count distribution
• score_vs_degree - Score vs network degree

generate_report()

generate_report(
    data_file: str,
    output_file: str,
    title: str = "Drug Discovery Analysis Report"
) -> str

Description: Generate comprehensive HTML report with all visualizations.

Parameters:

• data_file: Input data file path
• output_file: Output HTML file path
• title: Report title

Returns:

Path to generated HTML report