@undefineds.co/xpod 0.3.6 → 0.3.15

package/dist/storage/rdf/RdfTermDictionary.js

@@ -0,0 +1,352 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.RdfTermDictionary = void 0;
+const node_crypto_1 = require("node:crypto");
+const n3_1 = require("n3");
+const RdfTermSemantics_1 = require("./RdfTermSemantics");
+const XSD_STRING = 'http://www.w3.org/2001/XMLSchema#string';
+const RDF_LANG_STRING = 'http://www.w3.org/1999/02/22-rdf-syntax-ns#langString';
+class RdfTermDictionary {
+    constructor(db) {
+        this.db = db;
+        this.termCache = new Map();
+        this.idCache = new Map();
+    }
+    initialize() {
+        this.db.exec(`
+      CREATE TABLE IF NOT EXISTS rdf_terms (
+        id INTEGER PRIMARY KEY AUTOINCREMENT,
+        kind TEXT NOT NULL,
+        value TEXT NOT NULL,
+        datatype_id INTEGER,
+        lang TEXT,
+        hash TEXT NOT NULL,
+        normalized_text TEXT,
+        created_at TEXT NOT NULL DEFAULT (strftime('%Y-%m-%dT%H:%M:%fZ', 'now')),
+        UNIQUE (kind, value, datatype_id, lang)
+      );
+
+      CREATE INDEX IF NOT EXISTS rdf_terms_hash ON rdf_terms (hash);
+      CREATE INDEX IF NOT EXISTS rdf_terms_kind_value ON rdf_terms (kind, value);
+      CREATE INDEX IF NOT EXISTS rdf_terms_kind_datatype ON rdf_terms (kind, datatype_id);
+      CREATE INDEX IF NOT EXISTS rdf_terms_kind_lang ON rdf_terms (kind, lang);
+      CREATE INDEX IF NOT EXISTS rdf_terms_normalized_text ON rdf_terms (normalized_text);
+    `);
+        this.ensureNumericValueColumn();
+        this.db.exec('CREATE INDEX IF NOT EXISTS rdf_terms_kind_numeric_value ON rdf_terms (kind, numeric_value);');
+        this.backfillNumericValues();
+    }
+    getOrCreate(term) {
+        const identity = this.toIdentity(term);
+        const cacheKey = this.identityCacheKey(identity);
+        const cached = this.termCache.get(cacheKey);
+        if (cached !== undefined) {
+            return cached;
+        }
+        const existing = this.findId(identity);
+        if (existing !== undefined) {
+            this.termCache.set(cacheKey, existing);
+            return existing;
+        }
+        const stmt = this.db.prepare(`
+      INSERT INTO rdf_terms (kind, value, datatype_id, lang, hash, normalized_text, numeric_value)
+      VALUES (?, ?, ?, ?, ?, ?, ?)
+    `);
+        const result = stmt.run(identity.kind, identity.value, identity.datatypeId, identity.lang, identity.hash, identity.normalizedText, identity.numericValue);
+        const id = Number(result.lastInsertRowid);
+        this.termCache.set(cacheKey, id);
+        this.idCache.set(id, term);
+        return id;
+    }
+    find(term) {
+        const identity = this.toIdentity(term);
+        const cacheKey = this.identityCacheKey(identity);
+        const cached = this.termCache.get(cacheKey);
+        if (cached !== undefined) {
+            return cached;
+        }
+        const id = this.findId(identity);
+        if (id !== undefined) {
+            this.termCache.set(cacheKey, id);
+        }
+        return id;
+    }
+    termForId(id) {
+        const cached = this.idCache.get(id);
+        if (cached) {
+            return cached;
+        }
+        const row = this.db
+            .prepare('SELECT * FROM rdf_terms WHERE id = ?')
+            .get(id);
+        if (!row) {
+            throw new Error(`RDF term not found: ${id}`);
+        }
+        const term = this.rowToTerm(row);
+        this.idCache.set(id, term);
+        return term;
+    }
+    rowsForIds(ids) {
+        const uniqueIds = Array.from(new Set(ids));
+        const result = new Map();
+        const missing = [];
+        for (const id of uniqueIds) {
+            const cached = this.idCache.get(id);
+            if (cached) {
+                result.set(id, cached);
+            }
+            else {
+                missing.push(id);
+            }
+        }
+        if (missing.length === 0) {
+            return result;
+        }
+        const placeholders = missing.map(() => '?').join(', ');
+        const rows = this.db
+            .prepare(`SELECT * FROM rdf_terms WHERE id IN (${placeholders})`)
+            .all(...missing);
+        for (const row of rows) {
+            const term = this.rowToTerm(row);
+            this.idCache.set(row.id, term);
+            result.set(row.id, term);
+        }
+        return result;
+    }
+    idsByValuePrefix(kind, prefix) {
+        const kinds = Array.isArray(kind) ? kind : [kind];
+        if (kinds.length === 0) {
+            return [];
+        }
+        const rows = this.db
+            .prepare(`
+        SELECT id FROM rdf_terms
+        WHERE kind IN (${kinds.map(() => '?').join(', ')})
+          AND value >= ?
+          AND value < ?
+      `)
+            .all(...kinds, prefix, `${prefix}\uffff`);
+        return rows.map((row) => row.id);
+    }
+    idsByNormalizedTextContains(kind, text) {
+        const kinds = Array.isArray(kind) ? kind : [kind];
+        if (kinds.length === 0) {
+            return [];
+        }
+        const needle = normalizeSearchText(text);
+        const rows = this.db
+            .prepare(`
+        SELECT id, value FROM rdf_terms
+        WHERE kind IN (${kinds.map(() => '?').join(', ')})
+          AND normalized_text LIKE ? ESCAPE '\\'
+      `)
+            .all(...kinds, `%${escapeLikePattern(needle)}%`);
+        return rows
+            .filter((row) => row.value.includes(text))
+            .map((row) => row.id);
+    }
+    idsByNormalizedTextSuffix(kind, suffix) {
+        const kinds = Array.isArray(kind) ? kind : [kind];
+        if (kinds.length === 0) {
+            return [];
+        }
+        const needle = normalizeSearchText(suffix);
+        const rows = this.db
+            .prepare(`
+        SELECT id, value FROM rdf_terms
+        WHERE kind IN (${kinds.map(() => '?').join(', ')})
+          AND normalized_text LIKE ? ESCAPE '\\'
+      `)
+            .all(...kinds, `%${escapeLikePattern(needle)}`);
+        return rows
+            .filter((row) => row.value.endsWith(suffix))
+            .map((row) => row.id);
+    }
+    idsByNormalizedTextRegex(kind, pattern, flags) {
+        const kinds = Array.isArray(kind) ? kind : [kind];
+        if (kinds.length === 0) {
+            return [];
+        }
+        const rows = this.db
+            .prepare(`
+        SELECT * FROM rdf_terms
+        WHERE kind IN (${kinds.map(() => '?').join(', ')})
+          AND normalized_text IS NOT NULL
+      `)
+            .all(...kinds);
+        const regex = new RegExp(pattern, flags);
+        return rows
+            .filter((row) => regex.test(row.value))
+            .map((row) => row.id);
+    }
+    count() {
+        const row = this.db.prepare('SELECT COUNT(*) AS count FROM rdf_terms').get();
+        return row?.count ?? 0;
+    }
+    findId(identity) {
+        const row = identity.datatypeId === null && identity.lang === null
+            ? this.db
+                .prepare(`
+            SELECT id FROM rdf_terms
+            WHERE kind = ?
+              AND value = ?
+              AND datatype_id IS NULL
+              AND lang IS NULL
+            LIMIT 1
+          `)
+                .get(identity.kind, identity.value)
+            : identity.datatypeId === null
+                ? this.db
+                    .prepare(`
+              SELECT id FROM rdf_terms
+              WHERE kind = ?
+                AND value = ?
+                AND datatype_id IS NULL
+                AND lang = ?
+              LIMIT 1
+            `)
+                    .get(identity.kind, identity.value, identity.lang)
+                : identity.lang === null
+                    ? this.db
+                        .prepare(`
+                SELECT id FROM rdf_terms
+                WHERE kind = ?
+                  AND value = ?
+                  AND datatype_id = ?
+                  AND lang IS NULL
+                LIMIT 1
+              `)
+                        .get(identity.kind, identity.value, identity.datatypeId)
+                    : this.db
+                        .prepare(`
+                SELECT id FROM rdf_terms
+                WHERE kind = ?
+                  AND value = ?
+                  AND datatype_id = ?
+                  AND lang = ?
+                LIMIT 1
+              `)
+                        .get(identity.kind, identity.value, identity.datatypeId, identity.lang);
+        return row?.id;
+    }
+    toIdentity(term) {
+        switch (term.termType) {
+            case 'NamedNode':
+                return this.identity('iri', term.value, null, null, term.value, null);
+            case 'BlankNode':
+                return this.identity('blank', term.value, null, null, term.value, null);
+            case 'DefaultGraph':
+                return this.identity('default_graph', '', null, null, null, null);
+            case 'Literal': {
+                const datatypeValue = term.datatype?.value || XSD_STRING;
+                const datatypeId = datatypeValue === XSD_STRING && !term.language
+                    ? null
+                    : this.getOrCreate(n3_1.DataFactory.namedNode(datatypeValue));
+                return this.identity('literal', term.value, datatypeId, term.language || null, term.value, this.numericValueForLiteral(term.value, datatypeValue));
+            }
+            case 'Variable':
+                throw new Error(`Variables cannot be indexed as RDF terms: ${term.value}`);
+            case 'Quad':
+                throw new Error('Nested RDF-star quads are not supported by the first SolidRdfEngine index');
+            default: {
+                const exhaustive = term;
+                throw new Error(`Unsupported RDF term: ${String(exhaustive)}`);
+            }
+        }
+    }
+    identity(kind, value, datatypeId, lang, normalizedText, numericValue) {
+        const hash = (0, node_crypto_1.createHash)('sha256')
+            .update(kind)
+            .update('\0')
+            .update(value)
+            .update('\0')
+            .update(String(datatypeId ?? ''))
+            .update('\0')
+            .update(lang ?? '')
+            .digest('hex');
+        return {
+            kind,
+            value,
+            datatypeId,
+            lang,
+            normalizedText: normalizedText ? normalizedText.toLowerCase() : null,
+            numericValue,
+            hash,
+        };
+    }
+    numericValueForLiteral(value, datatypeValue) {
+        if (!(0, RdfTermSemantics_1.isRdfNumericDatatype)(datatypeValue)) {
+            return null;
+        }
+        const numeric = (0, RdfTermSemantics_1.rdfNumericValue)(value);
+        return Number.isFinite(numeric) ? numeric : null;
+    }
+    ensureNumericValueColumn() {
+        const columns = this.db.prepare('PRAGMA table_info(rdf_terms)').all();
+        if (columns.some((column) => column.name === 'numeric_value')) {
+            return;
+        }
+        this.db.exec('ALTER TABLE rdf_terms ADD COLUMN numeric_value REAL;');
+    }
+    backfillNumericValues() {
+        const rows = this.db.prepare(`
+      SELECT literal.*
+      FROM rdf_terms literal
+      JOIN rdf_terms datatype ON datatype.id = literal.datatype_id
+      WHERE literal.kind = 'literal'
+        AND literal.numeric_value IS NULL
+        AND datatype.kind = 'iri'
+    `).all();
+        const update = this.db.prepare('UPDATE rdf_terms SET numeric_value = ? WHERE id = ?');
+        for (const row of rows) {
+            const datatype = this.termForId(row.datatype_id);
+            if (datatype.termType !== 'NamedNode') {
+                continue;
+            }
+            const numericValue = this.numericValueForLiteral(row.value, datatype.value);
+            if (numericValue !== null) {
+                update.run(numericValue, row.id);
+            }
+        }
+    }
+    identityCacheKey(identity) {
+        return [
+            identity.kind,
+            identity.value,
+            identity.datatypeId ?? '',
+            identity.lang ?? '',
+        ].join('\u001f');
+    }
+    rowToTerm(row) {
+        switch (row.kind) {
+            case 'iri':
+                return n3_1.DataFactory.namedNode(row.value);
+            case 'blank':
+                return n3_1.DataFactory.blankNode(row.value);
+            case 'default_graph':
+                return n3_1.DataFactory.defaultGraph();
+            case 'literal': {
+                if (row.lang) {
+                    return n3_1.DataFactory.literal(row.value, row.lang);
+                }
+                if (row.datatype_id) {
+                    const datatype = this.termForId(row.datatype_id);
+                    if (datatype.termType === 'NamedNode' && datatype.value !== XSD_STRING && datatype.value !== RDF_LANG_STRING) {
+                        return n3_1.DataFactory.literal(row.value, datatype);
+                    }
+                }
+                return n3_1.DataFactory.literal(row.value);
+            }
+            default:
+                throw new Error(`Unsupported RDF term kind in dictionary row: ${row.kind}`);
+        }
+    }
+}
+exports.RdfTermDictionary = RdfTermDictionary;
+function normalizeSearchText(value) {
+    return value.toLowerCase();
+}
+function escapeLikePattern(value) {
+    return value.replace(/[\\%_]/g, (match) => `\\${match}`);
+}
+//# sourceMappingURL=RdfTermDictionary.js.map

package/dist/storage/rdf/RdfTermDictionary.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"RdfTermDictionary.js","sourceRoot":"","sources":["../../../src/storage/rdf/RdfTermDictionary.ts"],"names":[],"mappings":";;;AAAA,6CAAyC;AACzC,2BAAiC;AAIjC,yDAA2E;AAY3E,MAAM,UAAU,GAAG,yCAAyC,CAAC;AAC7D,MAAM,eAAe,GAAG,uDAAuD,CAAC;AAEhF,MAAa,iBAAiB;IAI5B,YAAoC,EAAkB;QAAlB,OAAE,GAAF,EAAE,CAAgB;QAHrC,cAAS,GAAG,IAAI,GAAG,EAAkB,CAAC;QACtC,YAAO,GAAG,IAAI,GAAG,EAAgB,CAAC;IAEM,CAAC;IAEnD,UAAU;QACf,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC;;;;;;;;;;;;;;;;;;KAkBZ,CAAC,CAAC;QACH,IAAI,CAAC,wBAAwB,EAAE,CAAC;QAChC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,6FAA6F,CAAC,CAAC;QAC5G,IAAI,CAAC,qBAAqB,EAAE,CAAC;IAC/B,CAAC;IAEM,WAAW,CAAC,IAAU;QAC3B,MAAM,QAAQ,GAAG,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC;QACvC,MAAM,QAAQ,GAAG,IAAI,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAC;QACjD,MAAM,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QAC5C,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;YACzB,OAAO,MAAM,CAAC;QAChB,CAAC;QAED,MAAM,QAAQ,GAAG,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;QACvC,IAAI,QAAQ,KAAK,SAAS,EAAE,CAAC;YAC3B,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC;YACvC,OAAO,QAAQ,CAAC;QAClB,CAAC;QAED,MAAM,IAAI,GAAG,IAAI,CAAC,EAAE,CAAC,OAAO,CAAC;;;KAG5B,CAAC,CAAC;QACH,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CACrB,QAAQ,CAAC,IAAI,EACb,QAAQ,CAAC,KAAK,EACd,QAAQ,CAAC,UAAU,EACnB,QAAQ,CAAC,IAAI,EACb,QAAQ,CAAC,IAAI,EACb,QAAQ,CAAC,cAAc,EACvB,QAAQ,CAAC,YAAY,CACtB,CAAC;QACF,MAAM,EAAE,GAAG,MAAM,CAAC,MAAM,CAAC,eAAe,CAAC,CAAC;QAC1C,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;QACjC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC;QAC3B,OAAO,EAAE,CAAC;IACZ,CAAC;IAEM,IAAI,CAAC,IAAU;QACpB,MAAM,QAAQ,GAAG,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC;QACvC,MAAM,QAAQ,GAAG,IAAI,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAC;QACjD,MAAM,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QAC5C,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;YACzB,OAAO,MAAM,CAAC;QAChB,CAAC;QAED,MAAM,EAAE,GAAG,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;QACjC,IAAI,EAAE,KAAK,SAAS,EAAE,CAAC;YACrB,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;QACnC,CAAC;QACD,OAAO,EAAE,CAAC;IACZ,CAAC;IAEM,SAAS,CAAC,EAAU;QACzB,MAAM,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QACpC,IAAI,MAAM,EAAE,CAAC;YACX,OAAO,MAAM,CAAC;QAChB,CAAC;QAED,MAAM,GAAG,GAAG,IAAI,CAAC,EAAE;aAChB,OAAO,CAAa,sCAAsC,CAAC;aAC3D,GAAG,CAAC,EAAE,CAAC,CAAC;QACX,IAAI,CAAC,GAAG,EAAE,CAAC;YACT,MAAM,IAAI,KAAK,CAAC,uBAAuB,EAAE,EAAE,CAAC,CAAC;QAC/C,CAAC;QAED,MAAM,IAAI,GAAG,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC;QACjC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC;QAC3B,OAAO,IAAI,CAAC;IACd,CAAC;IAEM,UAAU,CAAC,GAAa;QAC7B,MAAM,SAAS,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC;QAC3C,MAAM,MAAM,GAAG,IAAI,GAAG,EAAgB,CAAC;QACvC,MAAM,OAAO,GAAa,EAAE,CAAC;QAE7B,KAAK,MAAM,EAAE,IAAI,SAAS,EAAE,CAAC;YAC3B,MAAM,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;YACpC,IAAI,MAAM,EAAE,CAAC;gBACX,MAAM,CAAC,GAAG,CAAC,EAAE,EAAE,MAAM,CAAC,CAAC;YACzB,CAAC;iBAAM,CAAC;gBACN,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YACnB,CAAC;QACH,CAAC;QAED,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACzB,OAAO,MAAM,CAAC;QAChB,CAAC;QAED,MAAM,YAAY,GAAG,OAAO,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACvD,MAAM,IAAI,GAAG,IAAI,CAAC,EAAE;aACjB,OAAO,CAAa,wCAAwC,YAAY,GAAG,CAAC;aAC5E,GAAG,CAAC,GAAG,OAAO,CAAC,CAAC;QAEnB,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;YACvB,MAAM,IAAI,GAAG,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC;YACjC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC;YAC/B,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC;QAC3B,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAEM,gBAAgB,CAAC,IAAiC,EAAE,MAAc;QACvE,MAAM,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QAClD,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACvB,OAAO,EAAE,CAAC;QACZ,CAAC;QACD,MAAM,IAAI,GAAG,IAAI,CAAC,EAAE;aACjB,OAAO,CAAiB;;yBAEN,KAAK,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC;;;OAGjD,CAAC;aACD,GAAG,CAAC,GAAG,KAAK,EAAE,MAAM,EAAE,GAAG,MAAM,QAAQ,CAAC,CAAC;QAC5C,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;IACnC,CAAC;IAEM,2BAA2B,CAAC,IAAiC,EAAE,IAAY;QAChF,MAAM,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QAClD,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACvB,OAAO,EAAE,CAAC;QACZ,CAAC;QACD,MAAM,MAAM,GAAG,mBAAmB,CAAC,IAAI,CAAC,CAAC;QACzC,MAAM,IAAI,GAAG,IAAI,CAAC,EAAE;aACjB,OAAO,CAAgC;;yBAErB,KAAK,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC;;OAEjD,CAAC;aACD,GAAG,CAAC,GAAG,KAAK,EAAE,IAAI,iBAAiB,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QACnD,OAAO,IAAI;aACR,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;aACzC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;IAC1B,CAAC;IAEM,yBAAyB,CAAC,IAAiC,EAAE,MAAc;QAChF,MAAM,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QAClD,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACvB,OAAO,EAAE,CAAC;QACZ,CAAC;QACD,MAAM,MAAM,GAAG,mBAAmB,CAAC,MAAM,CAAC,CAAC;QAC3C,MAAM,IAAI,GAAG,IAAI,CAAC,EAAE;aACjB,OAAO,CAAgC;;yBAErB,KAAK,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC;;OAEjD,CAAC;aACD,GAAG,CAAC,GAAG,KAAK,EAAE,IAAI,iBAAiB,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;QAClD,OAAO,IAAI;aACR,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;aAC3C,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;IAC1B,CAAC;IAEM,wBAAwB,CAAC,IAAiC,EAAE,OAAe,EAAE,KAAc;QAChG,MAAM,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QAClD,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACvB,OAAO,EAAE,CAAC;QACZ,CAAC;QACD,MAAM,IAAI,GAAG,IAAI,CAAC,EAAE;aACjB,OAAO,CAAa;;yBAEF,KAAK,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC;;OAEjD,CAAC;aACD,GAAG,CAAC,GAAG,KAAK,CAAC,CAAC;QACjB,MAAM,KAAK,GAAG,IAAI,MAAM,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;QACzC,OAAO,IAAI;aACR,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;aACtC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;IAC1B,CAAC;IAEM,KAAK;QACV,MAAM,GAAG,GAAG,IAAI,CAAC,EAAE,CAAC,OAAO,CAAoB,yCAAyC,CAAC,CAAC,GAAG,EAAE,CAAC;QAChG,OAAO,GAAG,EAAE,KAAK,IAAI,CAAC,CAAC;IACzB,CAAC;IAEO,MAAM,CAAC,QAAyB;QACtC,MAAM,GAAG,GAAG,QAAQ,CAAC,UAAU,KAAK,IAAI,IAAI,QAAQ,CAAC,IAAI,KAAK,IAAI;YAChE,CAAC,CAAC,IAAI,CAAC,EAAE;iBACJ,OAAO,CAAiB;;;;;;;WAOxB,CAAC;iBACD,GAAG,CAAC,QAAQ,CAAC,IAAI,EAAE,QAAQ,CAAC,KAAK,CAAC;YACvC,CAAC,CAAC,QAAQ,CAAC,UAAU,KAAK,IAAI;gBAC5B,CAAC,CAAC,IAAI,CAAC,EAAE;qBACJ,OAAO,CAAiB;;;;;;;aAOxB,CAAC;qBACD,GAAG,CAAC,QAAQ,CAAC,IAAI,EAAE,QAAQ,CAAC,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC;gBACtD,CAAC,CAAC,QAAQ,CAAC,IAAI,KAAK,IAAI;oBACtB,CAAC,CAAC,IAAI,CAAC,EAAE;yBACJ,OAAO,CAAiB;;;;;;;eAOxB,CAAC;yBACD,GAAG,CAAC,QAAQ,CAAC,IAAI,EAAE,QAAQ,CAAC,KAAK,EAAE,QAAQ,CAAC,UAAU,CAAC;oBAC5D,CAAC,CAAC,IAAI,CAAC,EAAE;yBACJ,OAAO,CAAiB;;;;;;;eAOxB,CAAC;yBACD,GAAG,CAAC,QAAQ,CAAC,IAAI,EAAE,QAAQ,CAAC,KAAK,EAAE,QAAQ,CAAC,UAAU,EAAE,QAAQ,CAAC,IAAI,CAAC,CAAC;QAClF,OAAO,GAAG,EAAE,EAAE,CAAC;IACjB,CAAC;IAEO,UAAU,CAAC,IAAU;QAC3B,QAAQ,IAAI,CAAC,QAAQ,EAAE,CAAC;YACtB,KAAK,WAAW;gBACd,OAAO,IAAI,CAAC,QAAQ,CAAC,KAAK,EAAE,IAAI,CAAC,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;YACxE,KAAK,WAAW;gBACd,OAAO,IAAI,CAAC,QAAQ,CAAC,OAAO,EAAE,IAAI,CAAC,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;YAC1E,KAAK,cAAc;gBACjB,OAAO,IAAI,CAAC,QAAQ,CAAC,eAAe,EAAE,EAAE,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC;YACpE,KAAK,SAAS,CAAC,CAAC,CAAC;gBACf,MAAM,aAAa,GAAG,IAAI,CAAC,QAAQ,EAAE,KAAK,IAAI,UAAU,CAAC;gBACzD,MAAM,UAAU,GAAG,aAAa,KAAK,UAAU,IAAI,CAAC,IAAI,CAAC,QAAQ;oBAC/D,CAAC,CAAC,IAAI;oBACN,CAAC,CAAC,IAAI,CAAC,WAAW,CAAC,gBAAW,CAAC,SAAS,CAAC,aAAa,CAAC,CAAC,CAAC;gBAC3D,OAAO,IAAI,CAAC,QAAQ,CAClB,SAAS,EACT,IAAI,CAAC,KAAK,EACV,UAAU,EACV,IAAI,CAAC,QAAQ,IAAI,IAAI,EACrB,IAAI,CAAC,KAAK,EACV,IAAI,CAAC,sBAAsB,CAAC,IAAI,CAAC,KAAK,EAAE,aAAa,CAAC,CACvD,CAAC;YACJ,CAAC;YACD,KAAK,UAAU;gBACb,MAAM,IAAI,KAAK,CAAC,6CAA6C,IAAI,CAAC,KAAK,EAAE,CAAC,CAAC;YAC7E,KAAK,MAAM;gBACT,MAAM,IAAI,KAAK,CAAC,2EAA2E,CAAC,CAAC;YAC/F,OAAO,CAAC,CAAC,CAAC;gBACR,MAAM,UAAU,GAAU,IAAI,CAAC;gBAC/B,MAAM,IAAI,KAAK,CAAC,yBAAyB,MAAM,CAAC,UAAU,CAAC,EAAE,CAAC,CAAC;YACjE,CAAC;QACH,CAAC;IACH,CAAC;IAEO,QAAQ,CACd,IAAiB,EACjB,KAAa,EACb,UAAyB,EACzB,IAAmB,EACnB,cAA6B,EAC7B,YAA2B;QAE3B,MAAM,IAAI,GAAG,IAAA,wBAAU,EAAC,QAAQ,CAAC;aAC9B,MAAM,CAAC,IAAI,CAAC;aACZ,MAAM,CAAC,IAAI,CAAC;aACZ,MAAM,CAAC,KAAK,CAAC;aACb,MAAM,CAAC,IAAI,CAAC;aACZ,MAAM,CAAC,MAAM,CAAC,UAAU,IAAI,EAAE,CAAC,CAAC;aAChC,MAAM,CAAC,IAAI,CAAC;aACZ,MAAM,CAAC,IAAI,IAAI,EAAE,CAAC;aAClB,MAAM,CAAC,KAAK,CAAC,CAAC;QACjB,OAAO;YACL,IAAI;YACJ,KAAK;YACL,UAAU;YACV,IAAI;YACJ,cAAc,EAAE,cAAc,CAAC,CAAC,CAAC,cAAc,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,IAAI;YACpE,YAAY;YACZ,IAAI;SACL,CAAC;IACJ,CAAC;IAEO,sBAAsB,CAAC,KAAa,EAAE,aAAqB;QACjE,IAAI,CAAC,IAAA,uCAAoB,EAAC,aAAa,CAAC,EAAE,CAAC;YACzC,OAAO,IAAI,CAAC;QACd,CAAC;QACD,MAAM,OAAO,GAAG,IAAA,kCAAe,EAAC,KAAK,CAAC,CAAC;QACvC,OAAO,MAAM,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC;IACnD,CAAC;IAEO,wBAAwB;QAC9B,MAAM,OAAO,GAAG,IAAI,CAAC,EAAE,CAAC,OAAO,CAAmB,8BAA8B,CAAC,CAAC,GAAG,EAAE,CAAC;QACxF,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,KAAK,eAAe,CAAC,EAAE,CAAC;YAC9D,OAAO;QACT,CAAC;QACD,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,sDAAsD,CAAC,CAAC;IACvE,CAAC;IAEO,qBAAqB;QAC3B,MAAM,IAAI,GAAG,IAAI,CAAC,EAAE,CAAC,OAAO,CAAa;;;;;;;KAOxC,CAAC,CAAC,GAAG,EAAE,CAAC;QACT,MAAM,MAAM,GAAG,IAAI,CAAC,EAAE,CAAC,OAAO,CAAC,qDAAqD,CAAC,CAAC;QACtF,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;YACvB,MAAM,QAAQ,GAAG,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,WAAqB,CAAC,CAAC;YAC3D,IAAI,QAAQ,CAAC,QAAQ,KAAK,WAAW,EAAE,CAAC;gBACtC,SAAS;YACX,CAAC;YACD,MAAM,YAAY,GAAG,IAAI,CAAC,sBAAsB,CAAC,GAAG,CAAC,KAAK,EAAE,QAAQ,CAAC,KAAK,CAAC,CAAC;YAC5E,IAAI,YAAY,KAAK,IAAI,EAAE,CAAC;gBAC1B,MAAM,CAAC,GAAG,CAAC,YAAY,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC;YACnC,CAAC;QACH,CAAC;IACH,CAAC;IAEO,gBAAgB,CAAC,QAAyB;QAChD,OAAO;YACL,QAAQ,CAAC,IAAI;YACb,QAAQ,CAAC,KAAK;YACd,QAAQ,CAAC,UAAU,IAAI,EAAE;YACzB,QAAQ,CAAC,IAAI,IAAI,EAAE;SACpB,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;IACnB,CAAC;IAEO,SAAS,CAAC,GAAe;QAC/B,QAAQ,GAAG,CAAC,IAAI,EAAE,CAAC;YACjB,KAAK,KAAK;gBACR,OAAO,gBAAW,CAAC,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;YAC1C,KAAK,OAAO;gBACV,OAAO,gBAAW,CAAC,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;YAC1C,KAAK,eAAe;gBAClB,OAAO,gBAAW,CAAC,YAAY,EAAE,CAAC;YACpC,KAAK,SAAS,CAAC,CAAC,CAAC;gBACf,IAAI,GAAG,CAAC,IAAI,EAAE,CAAC;oBACb,OAAO,gBAAW,CAAC,OAAO,CAAC,GAAG,CAAC,KAAK,EAAE,GAAG,CAAC,IAAI,CAAC,CAAC;gBAClD,CAAC;gBACD,IAAI,GAAG,CAAC,WAAW,EAAE,CAAC;oBACpB,MAAM,QAAQ,GAAG,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC;oBACjD,IAAI,QAAQ,CAAC,QAAQ,KAAK,WAAW,IAAI,QAAQ,CAAC,KAAK,KAAK,UAAU,IAAI,QAAQ,CAAC,KAAK,KAAK,eAAe,EAAE,CAAC;wBAC7G,OAAO,gBAAW,CAAC,OAAO,CAAC,GAAG,CAAC,KAAK,EAAE,QAAQ,CAAC,CAAC;oBAClD,CAAC;gBACH,CAAC;gBACD,OAAO,gBAAW,CAAC,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;YACxC,CAAC;YACD;gBACE,MAAM,IAAI,KAAK,CAAC,gDAAiD,GAAkB,CAAC,IAAI,EAAE,CAAC,CAAC;QAChG,CAAC;IACH,CAAC;CACF;AA/XD,8CA+XC;AAED,SAAS,mBAAmB,CAAC,KAAa;IACxC,OAAO,KAAK,CAAC,WAAW,EAAE,CAAC;AAC7B,CAAC;AAED,SAAS,iBAAiB,CAAC,KAAa;IACtC,OAAO,KAAK,CAAC,OAAO,CAAC,SAAS,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,KAAK,EAAE,CAAC,CAAC;AAC3D,CAAC","sourcesContent":["import { createHash } from 'node:crypto';\nimport { DataFactory } from 'n3';\nimport type { Term } from '@rdfjs/types';\nimport type { SqliteDatabase } from '../SqliteRuntime';\nimport type { RdfTermKind, RdfTermRow } from './types';\nimport { isRdfNumericDatatype, rdfNumericValue } from './RdfTermSemantics';\n\ninterface RdfTermIdentity {\n  kind: RdfTermKind;\n  value: string;\n  datatypeId: number | null;\n  lang: string | null;\n  normalizedText: string | null;\n  numericValue: number | null;\n  hash: string;\n}\n\nconst XSD_STRING = 'http://www.w3.org/2001/XMLSchema#string';\nconst RDF_LANG_STRING = 'http://www.w3.org/1999/02/22-rdf-syntax-ns#langString';\n\nexport class RdfTermDictionary {\n  private readonly termCache = new Map<string, number>();\n  private readonly idCache = new Map<number, Term>();\n\n  public constructor(private readonly db: SqliteDatabase) {}\n\n  public initialize(): void {\n    this.db.exec(`\n      CREATE TABLE IF NOT EXISTS rdf_terms (\n        id INTEGER PRIMARY KEY AUTOINCREMENT,\n        kind TEXT NOT NULL,\n        value TEXT NOT NULL,\n        datatype_id INTEGER,\n        lang TEXT,\n        hash TEXT NOT NULL,\n        normalized_text TEXT,\n        created_at TEXT NOT NULL DEFAULT (strftime('%Y-%m-%dT%H:%M:%fZ', 'now')),\n        UNIQUE (kind, value, datatype_id, lang)\n      );\n\n      CREATE INDEX IF NOT EXISTS rdf_terms_hash ON rdf_terms (hash);\n      CREATE INDEX IF NOT EXISTS rdf_terms_kind_value ON rdf_terms (kind, value);\n      CREATE INDEX IF NOT EXISTS rdf_terms_kind_datatype ON rdf_terms (kind, datatype_id);\n      CREATE INDEX IF NOT EXISTS rdf_terms_kind_lang ON rdf_terms (kind, lang);\n      CREATE INDEX IF NOT EXISTS rdf_terms_normalized_text ON rdf_terms (normalized_text);\n    `);\n    this.ensureNumericValueColumn();\n    this.db.exec('CREATE INDEX IF NOT EXISTS rdf_terms_kind_numeric_value ON rdf_terms (kind, numeric_value);');\n    this.backfillNumericValues();\n  }\n\n  public getOrCreate(term: Term): number {\n    const identity = this.toIdentity(term);\n    const cacheKey = this.identityCacheKey(identity);\n    const cached = this.termCache.get(cacheKey);\n    if (cached !== undefined) {\n      return cached;\n    }\n\n    const existing = this.findId(identity);\n    if (existing !== undefined) {\n      this.termCache.set(cacheKey, existing);\n      return existing;\n    }\n\n    const stmt = this.db.prepare(`\n      INSERT INTO rdf_terms (kind, value, datatype_id, lang, hash, normalized_text, numeric_value)\n      VALUES (?, ?, ?, ?, ?, ?, ?)\n    `);\n    const result = stmt.run(\n      identity.kind,\n      identity.value,\n      identity.datatypeId,\n      identity.lang,\n      identity.hash,\n      identity.normalizedText,\n      identity.numericValue,\n    );\n    const id = Number(result.lastInsertRowid);\n    this.termCache.set(cacheKey, id);\n    this.idCache.set(id, term);\n    return id;\n  }\n\n  public find(term: Term): number | undefined {\n    const identity = this.toIdentity(term);\n    const cacheKey = this.identityCacheKey(identity);\n    const cached = this.termCache.get(cacheKey);\n    if (cached !== undefined) {\n      return cached;\n    }\n\n    const id = this.findId(identity);\n    if (id !== undefined) {\n      this.termCache.set(cacheKey, id);\n    }\n    return id;\n  }\n\n  public termForId(id: number): Term {\n    const cached = this.idCache.get(id);\n    if (cached) {\n      return cached;\n    }\n\n    const row = this.db\n      .prepare<RdfTermRow>('SELECT * FROM rdf_terms WHERE id = ?')\n      .get(id);\n    if (!row) {\n      throw new Error(`RDF term not found: ${id}`);\n    }\n\n    const term = this.rowToTerm(row);\n    this.idCache.set(id, term);\n    return term;\n  }\n\n  public rowsForIds(ids: number[]): Map<number, Term> {\n    const uniqueIds = Array.from(new Set(ids));\n    const result = new Map<number, Term>();\n    const missing: number[] = [];\n\n    for (const id of uniqueIds) {\n      const cached = this.idCache.get(id);\n      if (cached) {\n        result.set(id, cached);\n      } else {\n        missing.push(id);\n      }\n    }\n\n    if (missing.length === 0) {\n      return result;\n    }\n\n    const placeholders = missing.map(() => '?').join(', ');\n    const rows = this.db\n      .prepare<RdfTermRow>(`SELECT * FROM rdf_terms WHERE id IN (${placeholders})`)\n      .all(...missing);\n\n    for (const row of rows) {\n      const term = this.rowToTerm(row);\n      this.idCache.set(row.id, term);\n      result.set(row.id, term);\n    }\n\n    return result;\n  }\n\n  public idsByValuePrefix(kind: RdfTermKind | RdfTermKind[], prefix: string): number[] {\n    const kinds = Array.isArray(kind) ? kind : [kind];\n    if (kinds.length === 0) {\n      return [];\n    }\n    const rows = this.db\n      .prepare<{ id: number }>(`\n        SELECT id FROM rdf_terms\n        WHERE kind IN (${kinds.map(() => '?').join(', ')})\n          AND value >= ?\n          AND value < ?\n      `)\n      .all(...kinds, prefix, `${prefix}\\uffff`);\n    return rows.map((row) => row.id);\n  }\n\n  public idsByNormalizedTextContains(kind: RdfTermKind | RdfTermKind[], text: string): number[] {\n    const kinds = Array.isArray(kind) ? kind : [kind];\n    if (kinds.length === 0) {\n      return [];\n    }\n    const needle = normalizeSearchText(text);\n    const rows = this.db\n      .prepare<{ id: number; value: string }>(`\n        SELECT id, value FROM rdf_terms\n        WHERE kind IN (${kinds.map(() => '?').join(', ')})\n          AND normalized_text LIKE ? ESCAPE '\\\\'\n      `)\n      .all(...kinds, `%${escapeLikePattern(needle)}%`);\n    return rows\n      .filter((row) => row.value.includes(text))\n      .map((row) => row.id);\n  }\n\n  public idsByNormalizedTextSuffix(kind: RdfTermKind | RdfTermKind[], suffix: string): number[] {\n    const kinds = Array.isArray(kind) ? kind : [kind];\n    if (kinds.length === 0) {\n      return [];\n    }\n    const needle = normalizeSearchText(suffix);\n    const rows = this.db\n      .prepare<{ id: number; value: string }>(`\n        SELECT id, value FROM rdf_terms\n        WHERE kind IN (${kinds.map(() => '?').join(', ')})\n          AND normalized_text LIKE ? ESCAPE '\\\\'\n      `)\n      .all(...kinds, `%${escapeLikePattern(needle)}`);\n    return rows\n      .filter((row) => row.value.endsWith(suffix))\n      .map((row) => row.id);\n  }\n\n  public idsByNormalizedTextRegex(kind: RdfTermKind | RdfTermKind[], pattern: string, flags?: string): number[] {\n    const kinds = Array.isArray(kind) ? kind : [kind];\n    if (kinds.length === 0) {\n      return [];\n    }\n    const rows = this.db\n      .prepare<RdfTermRow>(`\n        SELECT * FROM rdf_terms\n        WHERE kind IN (${kinds.map(() => '?').join(', ')})\n          AND normalized_text IS NOT NULL\n      `)\n      .all(...kinds);\n    const regex = new RegExp(pattern, flags);\n    return rows\n      .filter((row) => regex.test(row.value))\n      .map((row) => row.id);\n  }\n\n  public count(): number {\n    const row = this.db.prepare<{ count: number }>('SELECT COUNT(*) AS count FROM rdf_terms').get();\n    return row?.count ?? 0;\n  }\n\n  private findId(identity: RdfTermIdentity): number | undefined {\n    const row = identity.datatypeId === null && identity.lang === null\n      ? this.db\n          .prepare<{ id: number }>(`\n            SELECT id FROM rdf_terms\n            WHERE kind = ?\n              AND value = ?\n              AND datatype_id IS NULL\n              AND lang IS NULL\n            LIMIT 1\n          `)\n          .get(identity.kind, identity.value)\n      : identity.datatypeId === null\n        ? this.db\n            .prepare<{ id: number }>(`\n              SELECT id FROM rdf_terms\n              WHERE kind = ?\n                AND value = ?\n                AND datatype_id IS NULL\n                AND lang = ?\n              LIMIT 1\n            `)\n            .get(identity.kind, identity.value, identity.lang)\n        : identity.lang === null\n          ? this.db\n              .prepare<{ id: number }>(`\n                SELECT id FROM rdf_terms\n                WHERE kind = ?\n                  AND value = ?\n                  AND datatype_id = ?\n                  AND lang IS NULL\n                LIMIT 1\n              `)\n              .get(identity.kind, identity.value, identity.datatypeId)\n          : this.db\n              .prepare<{ id: number }>(`\n                SELECT id FROM rdf_terms\n                WHERE kind = ?\n                  AND value = ?\n                  AND datatype_id = ?\n                  AND lang = ?\n                LIMIT 1\n              `)\n              .get(identity.kind, identity.value, identity.datatypeId, identity.lang);\n    return row?.id;\n  }\n\n  private toIdentity(term: Term): RdfTermIdentity {\n    switch (term.termType) {\n      case 'NamedNode':\n        return this.identity('iri', term.value, null, null, term.value, null);\n      case 'BlankNode':\n        return this.identity('blank', term.value, null, null, term.value, null);\n      case 'DefaultGraph':\n        return this.identity('default_graph', '', null, null, null, null);\n      case 'Literal': {\n        const datatypeValue = term.datatype?.value || XSD_STRING;\n        const datatypeId = datatypeValue === XSD_STRING && !term.language\n          ? null\n          : this.getOrCreate(DataFactory.namedNode(datatypeValue));\n        return this.identity(\n          'literal',\n          term.value,\n          datatypeId,\n          term.language || null,\n          term.value,\n          this.numericValueForLiteral(term.value, datatypeValue),\n        );\n      }\n      case 'Variable':\n        throw new Error(`Variables cannot be indexed as RDF terms: ${term.value}`);\n      case 'Quad':\n        throw new Error('Nested RDF-star quads are not supported by the first SolidRdfEngine index');\n      default: {\n        const exhaustive: never = term;\n        throw new Error(`Unsupported RDF term: ${String(exhaustive)}`);\n      }\n    }\n  }\n\n  private identity(\n    kind: RdfTermKind,\n    value: string,\n    datatypeId: number | null,\n    lang: string | null,\n    normalizedText: string | null,\n    numericValue: number | null,\n  ): RdfTermIdentity {\n    const hash = createHash('sha256')\n      .update(kind)\n      .update('\\0')\n      .update(value)\n      .update('\\0')\n      .update(String(datatypeId ?? ''))\n      .update('\\0')\n      .update(lang ?? '')\n      .digest('hex');\n    return {\n      kind,\n      value,\n      datatypeId,\n      lang,\n      normalizedText: normalizedText ? normalizedText.toLowerCase() : null,\n      numericValue,\n      hash,\n    };\n  }\n\n  private numericValueForLiteral(value: string, datatypeValue: string): number | null {\n    if (!isRdfNumericDatatype(datatypeValue)) {\n      return null;\n    }\n    const numeric = rdfNumericValue(value);\n    return Number.isFinite(numeric) ? numeric : null;\n  }\n\n  private ensureNumericValueColumn(): void {\n    const columns = this.db.prepare<{ name: string }>('PRAGMA table_info(rdf_terms)').all();\n    if (columns.some((column) => column.name === 'numeric_value')) {\n      return;\n    }\n    this.db.exec('ALTER TABLE rdf_terms ADD COLUMN numeric_value REAL;');\n  }\n\n  private backfillNumericValues(): void {\n    const rows = this.db.prepare<RdfTermRow>(`\n      SELECT literal.*\n      FROM rdf_terms literal\n      JOIN rdf_terms datatype ON datatype.id = literal.datatype_id\n      WHERE literal.kind = 'literal'\n        AND literal.numeric_value IS NULL\n        AND datatype.kind = 'iri'\n    `).all();\n    const update = this.db.prepare('UPDATE rdf_terms SET numeric_value = ? WHERE id = ?');\n    for (const row of rows) {\n      const datatype = this.termForId(row.datatype_id as number);\n      if (datatype.termType !== 'NamedNode') {\n        continue;\n      }\n      const numericValue = this.numericValueForLiteral(row.value, datatype.value);\n      if (numericValue !== null) {\n        update.run(numericValue, row.id);\n      }\n    }\n  }\n\n  private identityCacheKey(identity: RdfTermIdentity): string {\n    return [\n      identity.kind,\n      identity.value,\n      identity.datatypeId ?? '',\n      identity.lang ?? '',\n    ].join('\\u001f');\n  }\n\n  private rowToTerm(row: RdfTermRow): Term {\n    switch (row.kind) {\n      case 'iri':\n        return DataFactory.namedNode(row.value);\n      case 'blank':\n        return DataFactory.blankNode(row.value);\n      case 'default_graph':\n        return DataFactory.defaultGraph();\n      case 'literal': {\n        if (row.lang) {\n          return DataFactory.literal(row.value, row.lang);\n        }\n        if (row.datatype_id) {\n          const datatype = this.termForId(row.datatype_id);\n          if (datatype.termType === 'NamedNode' && datatype.value !== XSD_STRING && datatype.value !== RDF_LANG_STRING) {\n            return DataFactory.literal(row.value, datatype);\n          }\n        }\n        return DataFactory.literal(row.value);\n      }\n      default:\n        throw new Error(`Unsupported RDF term kind in dictionary row: ${(row as RdfTermRow).kind}`);\n    }\n  }\n}\n\nfunction normalizeSearchText(value: string): string {\n  return value.toLowerCase();\n}\n\nfunction escapeLikePattern(value: string): string {\n  return value.replace(/[\\\\%_]/g, (match) => `\\\\${match}`);\n}\n"]}

package/dist/storage/rdf/RdfTermDictionary.jsonld

@@ -0,0 +1,114 @@
+{
+  "@context": [
+    "https://linkedsoftwaredependencies.org/bundles/npm/@undefineds.co/xpod/^0.0.0/components/context.jsonld"
+  ],
+  "@id": "npmd:@undefineds.co/xpod",
+  "components": [
+    {
+      "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary",
+      "@type": "Class",
+      "requireElement": "RdfTermDictionary",
+      "parameters": [
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary_db",
+          "range": "undefineds:dist/storage/sqlite/types.jsonld#SqliteDatabase"
+        }
+      ],
+      "memberFields": [
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_db",
+          "memberFieldName": "db"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_termCache",
+          "memberFieldName": "termCache"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_idCache",
+          "memberFieldName": "idCache"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_constructor",
+          "memberFieldName": "constructor"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_initialize",
+          "memberFieldName": "initialize"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_getOrCreate",
+          "memberFieldName": "getOrCreate"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_find",
+          "memberFieldName": "find"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_termForId",
+          "memberFieldName": "termForId"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_rowsForIds",
+          "memberFieldName": "rowsForIds"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_idsByValuePrefix",
+          "memberFieldName": "idsByValuePrefix"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_idsByNormalizedTextContains",
+          "memberFieldName": "idsByNormalizedTextContains"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_idsByNormalizedTextSuffix",
+          "memberFieldName": "idsByNormalizedTextSuffix"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_idsByNormalizedTextRegex",
+          "memberFieldName": "idsByNormalizedTextRegex"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_count",
+          "memberFieldName": "count"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_findId",
+          "memberFieldName": "findId"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_toIdentity",
+          "memberFieldName": "toIdentity"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_identity",
+          "memberFieldName": "identity"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_numericValueForLiteral",
+          "memberFieldName": "numericValueForLiteral"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_ensureNumericValueColumn",
+          "memberFieldName": "ensureNumericValueColumn"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_backfillNumericValues",
+          "memberFieldName": "backfillNumericValues"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_identityCacheKey",
+          "memberFieldName": "identityCacheKey"
+        },
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary__member_rowToTerm",
+          "memberFieldName": "rowToTerm"
+        }
+      ],
+      "constructorArguments": [
+        {
+          "@id": "undefineds:dist/storage/rdf/RdfTermDictionary.jsonld#RdfTermDictionary_db"
+        }
+      ]
+    }
+  ]
+}

package/dist/storage/rdf/RdfTermSemantics.d.ts

@@ -0,0 +1,6 @@
+import type { Term } from '@rdfjs/types';
+export declare const RDF_NUMERIC_DATATYPES: readonly ["http://www.w3.org/2001/XMLSchema#integer", "http://www.w3.org/2001/XMLSchema#decimal", "http://www.w3.org/2001/XMLSchema#double", "http://www.w3.org/2001/XMLSchema#float", "http://www.w3.org/2001/XMLSchema#long", "http://www.w3.org/2001/XMLSchema#int", "http://www.w3.org/2001/XMLSchema#short", "http://www.w3.org/2001/XMLSchema#byte", "http://www.w3.org/2001/XMLSchema#nonNegativeInteger", "http://www.w3.org/2001/XMLSchema#nonPositiveInteger", "http://www.w3.org/2001/XMLSchema#positiveInteger", "http://www.w3.org/2001/XMLSchema#negativeInteger", "http://www.w3.org/2001/XMLSchema#unsignedLong", "http://www.w3.org/2001/XMLSchema#unsignedInt", "http://www.w3.org/2001/XMLSchema#unsignedShort", "http://www.w3.org/2001/XMLSchema#unsignedByte"];
+export declare function isRdfNumericDatatype(datatype: string): boolean;
+export declare function isFiniteNumericLexical(value: string | number): boolean;
+export declare function rdfNumericValue(value: string | number): number;
+export declare function isRdfNumericTerm(term: Term): boolean;

package/dist/storage/rdf/RdfTermSemantics.js

@@ -0,0 +1,40 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.RDF_NUMERIC_DATATYPES = void 0;
+exports.isRdfNumericDatatype = isRdfNumericDatatype;
+exports.isFiniteNumericLexical = isFiniteNumericLexical;
+exports.rdfNumericValue = rdfNumericValue;
+exports.isRdfNumericTerm = isRdfNumericTerm;
+exports.RDF_NUMERIC_DATATYPES = [
+    'http://www.w3.org/2001/XMLSchema#integer',
+    'http://www.w3.org/2001/XMLSchema#decimal',
+    'http://www.w3.org/2001/XMLSchema#double',
+    'http://www.w3.org/2001/XMLSchema#float',
+    'http://www.w3.org/2001/XMLSchema#long',
+    'http://www.w3.org/2001/XMLSchema#int',
+    'http://www.w3.org/2001/XMLSchema#short',
+    'http://www.w3.org/2001/XMLSchema#byte',
+    'http://www.w3.org/2001/XMLSchema#nonNegativeInteger',
+    'http://www.w3.org/2001/XMLSchema#nonPositiveInteger',
+    'http://www.w3.org/2001/XMLSchema#positiveInteger',
+    'http://www.w3.org/2001/XMLSchema#negativeInteger',
+    'http://www.w3.org/2001/XMLSchema#unsignedLong',
+    'http://www.w3.org/2001/XMLSchema#unsignedInt',
+    'http://www.w3.org/2001/XMLSchema#unsignedShort',
+    'http://www.w3.org/2001/XMLSchema#unsignedByte',
+];
+function isRdfNumericDatatype(datatype) {
+    return exports.RDF_NUMERIC_DATATYPES.includes(datatype);
+}
+function isFiniteNumericLexical(value) {
+    return Number.isFinite(Number(value));
+}
+function rdfNumericValue(value) {
+    return Number(value);
+}
+function isRdfNumericTerm(term) {
+    return term.termType === 'Literal'
+        && isRdfNumericDatatype(term.datatype.value)
+        && isFiniteNumericLexical(term.value);
+}
+//# sourceMappingURL=RdfTermSemantics.js.map

package/dist/storage/rdf/RdfTermSemantics.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"RdfTermSemantics.js","sourceRoot":"","sources":["../../../src/storage/rdf/RdfTermSemantics.ts"],"names":[],"mappings":";;;AAqBA,oDAEC;AAED,wDAEC;AAED,0CAEC;AAED,4CAIC;AAnCY,QAAA,qBAAqB,GAAG;IACnC,0CAA0C;IAC1C,0CAA0C;IAC1C,yCAAyC;IACzC,wCAAwC;IACxC,uCAAuC;IACvC,sCAAsC;IACtC,wCAAwC;IACxC,uCAAuC;IACvC,qDAAqD;IACrD,qDAAqD;IACrD,kDAAkD;IAClD,kDAAkD;IAClD,+CAA+C;IAC/C,8CAA8C;IAC9C,gDAAgD;IAChD,+CAA+C;CACvC,CAAC;AAEX,SAAgB,oBAAoB,CAAC,QAAgB;IACnD,OAAO,6BAAqB,CAAC,QAAQ,CAAC,QAAgD,CAAC,CAAC;AAC1F,CAAC;AAED,SAAgB,sBAAsB,CAAC,KAAsB;IAC3D,OAAO,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC;AACxC,CAAC;AAED,SAAgB,eAAe,CAAC,KAAsB;IACpD,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC;AACvB,CAAC;AAED,SAAgB,gBAAgB,CAAC,IAAU;IACzC,OAAO,IAAI,CAAC,QAAQ,KAAK,SAAS;WAC7B,oBAAoB,CAAC,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC;WACzC,sBAAsB,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;AAC1C,CAAC","sourcesContent":["import type { Term } from '@rdfjs/types';\n\nexport const RDF_NUMERIC_DATATYPES = [\n  'http://www.w3.org/2001/XMLSchema#integer',\n  'http://www.w3.org/2001/XMLSchema#decimal',\n  'http://www.w3.org/2001/XMLSchema#double',\n  'http://www.w3.org/2001/XMLSchema#float',\n  'http://www.w3.org/2001/XMLSchema#long',\n  'http://www.w3.org/2001/XMLSchema#int',\n  'http://www.w3.org/2001/XMLSchema#short',\n  'http://www.w3.org/2001/XMLSchema#byte',\n  'http://www.w3.org/2001/XMLSchema#nonNegativeInteger',\n  'http://www.w3.org/2001/XMLSchema#nonPositiveInteger',\n  'http://www.w3.org/2001/XMLSchema#positiveInteger',\n  'http://www.w3.org/2001/XMLSchema#negativeInteger',\n  'http://www.w3.org/2001/XMLSchema#unsignedLong',\n  'http://www.w3.org/2001/XMLSchema#unsignedInt',\n  'http://www.w3.org/2001/XMLSchema#unsignedShort',\n  'http://www.w3.org/2001/XMLSchema#unsignedByte',\n] as const;\n\nexport function isRdfNumericDatatype(datatype: string): boolean {\n  return RDF_NUMERIC_DATATYPES.includes(datatype as typeof RDF_NUMERIC_DATATYPES[number]);\n}\n\nexport function isFiniteNumericLexical(value: string | number): boolean {\n  return Number.isFinite(Number(value));\n}\n\nexport function rdfNumericValue(value: string | number): number {\n  return Number(value);\n}\n\nexport function isRdfNumericTerm(term: Term): boolean {\n  return term.termType === 'Literal'\n    && isRdfNumericDatatype(term.datatype.value)\n    && isFiniteNumericLexical(term.value);\n}\n"]}

package/dist/storage/rdf/RdfTextIndex.d.ts

@@ -0,0 +1,23 @@
+import type { RdfTextChunkInput, RdfTextIndexOptions, RdfTextIndexStats, RdfSearchCardinalityEstimate, RdfTextSearchOptions, RdfTextSearchResult, RdfTextSourceInput, RdfTextTermDocumentFrequency } from './types';
+export declare class RdfTextIndex {
+    private readonly options;
+    private readonly sqliteRuntime;
+    private db;
+    constructor(options: RdfTextIndexOptions);
+    open(): void;
+    close(): void;
+    clear(): void;
+    indexText(source: RdfTextSourceInput, text: string, chunks?: RdfTextChunkInput[]): void;
+    deleteSource(source: string): number;
+    search(options: RdfTextSearchOptions): RdfTextSearchResult[];
+    estimateSearchCardinality(options: RdfTextSearchOptions): RdfSearchCardinalityEstimate;
+    stats(): RdfTextIndexStats;
+    termDocumentFrequency(limit?: number): RdfTextTermDocumentFrequency[];
+    private initializeSchema;
+    private backfillTermPostings;
+    private upsertSource;
+    private chunkText;
+    private estimateDatabaseBytes;
+    private toSearchResult;
+    private requireDb;
+}