add compound metadata

solr_metadata.py

@@ -358,6 +358,286 @@ field_metadata = [
         "type": "string (multi-valued, categorical)",
         "example_values": ["Small Molecules", "Biologics", "Nucleic Acids", "Peptides"],
         "definition": "High-level classification of the drug's molecular type involved in the deal."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "id",
+        "type": "string",
+        "example_values": ["519-571", "35518_child"],
+        "definition": "Unique identifier for the compound document."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "molecule_id",
+        "type": "string",
+        "example_values": ["mol_12345", "mol_67890"], # Placeholder, not in provided doc examples directly.
+        "definition": "Unique identifier for the molecule associated with the compound."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "product_name",
+        "type": "string (exact match, for faceting)",
+        "example_values": ["Toprol-XL Tablets"],
+        "definition": "The specific, full trade name of the product. **Use this field for `terms` faceting** on products."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "product_name_s",
+        "type": "string (multi-valued, for searching)",
+        "example_values": ["Toprol-XL Tablets", "Toprol Extended Release Tablets"],
+        "definition": "A field containing all known trade names and synonyms for a product. **Use this field for all `query` parameter searches involving a product name** to ensure comprehensive results."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "product_synonym",
+        "type": "string (multi-valued)",
+        "example_values": ["Toprol Extended Release Tablets", "metoprolol succinate ER Tablets"],
+        "definition": "Synonyms or alternative names for the product."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "molecule_name",
+        "type": "string (exact match, for faceting)",
+        "example_values": ["metoprolol succinate"],
+        "definition": "The generic, non-proprietary name of the active molecule. **Use this field for `terms` faceting** on molecules."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "molecule_name_s",
+        "type": "string (multi-valued, for searching)",
+        "example_values": ["metoprolol succinate", "Métoprolol, Succinate de"],
+        "definition": "A field with all known generic names and synonyms for a molecule. **Use this field for all `query` parameter searches** involving a molecule name."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "molecule_synonym",
+        "type": "string (multi-valued)",
+        "example_values": ["Métoprolol, Succinate de", "Metoprololi Succinas"],
+        "definition": "Synonyms or alternative names for the molecule."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "molecule_type",
+        "type": "string (categorical)",
+        "example_values": ["small molecule", "biologic", "nucleic acid"],
+        "definition": "High-level classification of the compound's molecular type."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "company_name",
+        "type": "string (multi-valued, exact match, for faceting)",
+        "example_values": ["AstraZeneca Plc", "Recordati S.p.A."],
+        "definition": "The canonical, standardized names of companies associated with the compound. **Use this field for `terms` faceting** to group results by unique companies."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "company_name_s",
+        "type": "string (multi-valued, for searching)",
+        "example_values": ["AstraZeneca Plc", "New American Therapeutics, Inc."],
+        "definition": "A field containing all known names and synonyms for companies associated with the compound. **Use this field for all `query` parameter searches involving a company name**."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "company_role",
+        "type": "string (multi-valued, categorical)",
+        "example_values": ["owner", "partner", "manufacturer"],
+        "definition": "The role of the company in relation to the compound (e.g., owner, partner, manufacturer)."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "owner_company_name",
+        "type": "string (multi-valued, exact match, for faceting)",
+        "example_values": ["AstraZeneca Plc"], # Placeholder, based on common patterns
+        "definition": "The canonical, standardized name(s) of the owner company of the compound. Use for faceting on owner companies."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "overall_status",
+        "type": "string (multi-valued, categorical)",
+        "example_values": ["Marketed", "Completed", "Terminated"],
+        "definition": "The overall development status of the compound (e.g., Marketed, Completed, Terminated)."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "phase",
+        "type": "string (multi-valued, categorical)",
+        "example_values": ["Marketed", "Phase 1", "Pre Clinical"],
+        "definition": "The current or most advanced clinical or developmental stage of the compound. Essential for queries about clinical trial phases."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "highest_phase_molecule",
+        "type": "string (multi-valued, categorical)",
+        "example_values": ["Marketed", "Phase 3"],
+        "definition": "The highest development stage a molecule associated with this compound has ever reached."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "therapeutic_category",
+        "type": "string (multi-valued, specific, for faceting)",
+        "example_values": ["Angina Pectoris", "Heart Failure, Other"],
+        "definition": "The specific disease or therapeutic area being targeted by the compound. Use for very specific disease queries or faceting."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "therapeutic_category_s",
+        "type": "string (multi-valued, for searching)",
+        "example_values": ["cardiovascular diseases", "heart diseases", "angina pectoris"],
+        "definition": "Broader, multi-valued therapeutic categories and their synonyms. **Use this field for broad category searches** in the `query` parameter."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "therapeutic_category_hierarchy",
+        "type": "string (multi-valued, hierarchical)",
+        "example_values": ["Cardiovascular Diseases|Heart Diseases|Coronary Artery Disease|Angina Pectoris"],
+        "definition": "The hierarchical classification of the therapeutic area (e.g., 'Category|Subcategory|Specific Area'). Useful for hierarchical faceting or broader path-based searches. [12, 16, 18, 24]"
+    },
+    {
+        "core_name": "compound",
+        "field_name": "mechanism",
+        "type": "string (multi-valued, specific, for faceting)",
+        "example_values": ["Beta-1 Adrenergic Receptor Antagonist"], # Placeholder
+        "definition": "The specific mechanism of action of the compound. Use for very specific mechanism queries or faceting."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "mechanism_s",
+        "type": "string (multi-valued, for searching)",
+        "example_values": ["beta blocker", "adrenergic receptor antagonist"], # Placeholder
+        "definition": "Broader, multi-valued mechanisms of action and their synonyms. **Use this field for broad mechanism searches** in the `query` parameter."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "mechanism_type",
+        "type": "string (multi-valued, categorical)",
+        "example_values": ["Cardioprotectants", "Antihypertensive Agents", "Beta-1 Blockers"],
+        "definition": "High-level classification of the mechanism type."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "mechanism_type_hierarchy",
+        "type": "string (multi-valued, hierarchical)",
+        "example_values": ["Tissue Protectants|Cardioprotectants", "Adrenergic Receptor Modulators|Beta Adrenergic Receptor Modulators|Beta Blockers|Beta-1 Blockers"],
+        "definition": "The hierarchical classification of the mechanism type."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "target_name",
+        "type": "string (multi-valued, exact match, for faceting)",
+        "example_values": ["Beta1 Adrenergic"],
+        "definition": "The specific name of the biological target(s) of the compound. **Use this field for `terms` faceting** on targets."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "target_name_s",
+        "type": "string (multi-valued, for searching)",
+        "example_values": ["Beta1 Adrenergic", "Beta-1 Adrenergic Receptor", "ADRB1"],
+        "definition": "A field with all known names and synonyms for the biological target(s). **Use this field for all `query` parameter searches** involving a target name."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "target_type",
+        "type": "string (multi-valued, categorical)",
+        "example_values": ["protein", "biochemical process/pathway", "other"],
+        "definition": "The type of biological target (e.g., protein, pathway)."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "route",
+        "type": "string (multi-valued, categorical)",
+        "example_values": ["ORAL", "INJECTION", "TOPICAL"],
+        "definition": "The primary route of administration for the compound. Good for faceting on exact routes."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "route_s",
+        "type": "string (multi-valued, for searching)",
+        "example_values": ["oral", "parenteral", "injection"],
+        "definition": "Broader, multi-valued routes of administration and their synonyms. **Use this field for broad route searches** in the `query` parameter."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "route_branch_hierarchy",
+        "type": "string (multi-valued, hierarchical)",
+        "example_values": ["ORAL"],
+        "definition": "The hierarchical classification of the route of administration."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "form_name",
+        "type": "string (multi-valued, categorical)",
+        "example_values": ["ORAL|Oral Tablet", "INJECTION|Solution"],
+        "definition": "The pharmaceutical form of the compound, often combined with the route (e.g., 'Route|Form')."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "drug_delivery_technology",
+        "type": "string (multi-valued, specific, for faceting)",
+        "example_values": ["Conventional Melt Tablets", "Oral Matrix MR", "Bioadhesion"],
+        "definition": "Specific categories of drug delivery technology employed for the compound. **Use this field for `terms` faceting** on specific delivery technologies."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "drug_delivery_branch_s",
+        "type": "string (multi-valued, for searching)",
+        "example_values": ["oral", "oral modified release", "oral extended release"],
+        "definition": "The method of drug administration technology and its broader search-friendly terms. **Use this for `query` parameter searches about drug delivery technologies**."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "earliest_approval_date",
+        "type": "date",
+        "example_values": ["1986-06-19T00:00:00Z"],
+        "definition": "The earliest known approval date for the compound in ISO 8601 format. Use for precise date range queries."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "approval_year",
+        "type": "number (multi-valued, year)",
+        "example_values": [1992, 2000, 2001],
+        "definition": "The 4-digit year(s) of approval for the compound. Use for queries involving whole years."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "compound_territory",
+        "type": "string (multi-valued, hierarchical)",
+        "example_values": ["United States of America", "Americas Northern", "Europe", "World"],
+        "definition": "The geographic territories where the compound is approved or marketed. It is hierarchical. Use for filtering by location. [12, 16, 18, 24]"
+    },
+    {
+        "core_name": "compound",
+        "field_name": "is_orphan",
+        "type": "string (boolean)", # Solr typically stores booleans as strings or numbers [14, 15, 21, 23]
+        "example_values": ["yes", "no"],
+        "definition": "Indicates if the compound has received orphan drug designation. Orphan drugs are intended for rare diseases that affect a small population, making them less profitable to develop without incentives. [2, 4, 7, 9, 10]"
+    },
+    {
+        "core_name": "compound",
+        "field_name": "is_prodrug",
+        "type": "string (boolean)",
+        "example_values": ["yes", "no"],
+        "definition": "Indicates if the compound is a prodrug (an inactive compound that is metabolized in the body to an active drug)."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "black_box",
+        "type": "string (boolean)",
+        "example_values": ["yes", "no"],
+        "definition": "Indicates if the compound carries a 'Black Box Warning' (the strongest warning by the FDA that a drug carries significant risks)."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "molecular_weight",
+        "type": "number (metric)",
+        "example_values": [652.816, 300.25],
+        "definition": "The molecular weight of the compound in Daltons. Use for numerical range queries."
+    },
+    {
+        "core_name": "compound",
+        "field_name": "logP",
+        "type": "number (metric)",
+        "example_values": [2.5, -0.7, 4.1],
+        "definition": "The logarithm of the octanol-water partition coefficient, a measure of a compound's lipophilicity (fat-liking) or hydrophilicity (water-liking). Higher values indicate more lipophilicity. [1, 3, 5, 6, 8]"
     }
 ]