@opencodehub/ingestion 0.1.0

package/dist/pipeline/phases/scip-index.js

@@ -0,0 +1,469 @@
+/**
+ * `scip-index` phase — replaces the four per-language LSP upgrade
+ * phases (lsp-python / lsp-typescript / lsp-go / lsp-rust) with a
+ * single pass over SCIP (https://scip-code.org) indexes.
+ *
+ * For every language detected in the profile we:
+ *   1. Run the appropriate SCIP indexer into
+ *      `.codehub/scip/<lang>.scip`, unless the artifact is fresh
+ *      (mtime newer than every source file for that language).
+ *   2. Parse the index with `@opencodehub/scip-ingest`, derive caller
+ *      -> callee edges via innermost-enclosing-range attribution.
+ *   3. Map each SCIP call site (document, line) back to the tightest
+ *      OpenCodeHub symbol node via the same file+line lookup the LSP
+ *      phases used.
+ *   4. Emit CodeRelation edges with `confidence = 1.0` and
+ *      `reason = scip:<indexer>@<version>` so the downstream
+ *      `confidence-demote`, `summarize`, `mcp/confidence`, and
+ *      `cli/analyze` consumers keep treating SCIP edges as oracle-
+ *      confirmed (see `SCIP_PROVENANCE_PREFIXES`).
+ *
+ * Skip semantics:
+ *   - `CODEHUB_DISABLE_SCIP=1`       -> entire phase no-op.
+ *   - Indexer binary missing         -> per-language skip with a warn
+ *                                       ProgressEvent; tree-sitter
+ *                                       heuristic tier keeps its
+ *                                       low-confidence edges.
+ *   - `options.offline && !cached`   -> per-language skip (no network
+ *                                       install path).
+ */
+import { existsSync, statSync } from "node:fs";
+import { readFile } from "node:fs/promises";
+import { join } from "node:path";
+import { buildSymbolDefIndex, deriveIndex, detectLanguages, parseScipIndex, runIndexer, scipProvenanceReason, } from "@opencodehub/scip-ingest";
+import { META_DIR_NAME } from "@opencodehub/storage";
+import { ACCESSES_PHASE_NAME } from "./accesses.js";
+import { CROSS_FILE_PHASE_NAME } from "./cross-file.js";
+import { PARSE_PHASE_NAME } from "./parse.js";
+import { PROFILE_PHASE_NAME } from "./profile.js";
+import { SCAN_PHASE_NAME } from "./scan.js";
+export const SCIP_INDEX_PHASE_NAME = "scip-index";
+const SCIP_CONFIDENCE = 1.0;
+export const scipIndexPhase = {
+    name: SCIP_INDEX_PHASE_NAME,
+    deps: [
+        SCAN_PHASE_NAME,
+        PROFILE_PHASE_NAME,
+        PARSE_PHASE_NAME,
+        CROSS_FILE_PHASE_NAME,
+        ACCESSES_PHASE_NAME,
+    ],
+    async run(ctx, deps) {
+        return runScipIndex(ctx, deps);
+    },
+};
+async function runScipIndex(ctx, deps) {
+    const start = Date.now();
+    if (process.env["CODEHUB_DISABLE_SCIP"] === "1") {
+        return {
+            enabled: false,
+            skippedReason: "CODEHUB_DISABLE_SCIP=1",
+            languages: [],
+            totalEdgesAdded: 0,
+            totalEdgesUpgraded: 0,
+            durationMs: Date.now() - start,
+        };
+    }
+    const profile = deps.get(PROFILE_PHASE_NAME);
+    const profileNode = findProfileNode(ctx);
+    if (profile === undefined || profileNode === undefined) {
+        return {
+            enabled: false,
+            skippedReason: "no-profile-output",
+            languages: [],
+            totalEdgesAdded: 0,
+            totalEdgesUpgraded: 0,
+            durationMs: Date.now() - start,
+        };
+    }
+    const projectLanguages = new Set(profileNode.languages);
+    const candidates = detectLanguages(ctx.repoPath).filter((k) => projectLanguages.has(scipLangToOchLang(k)));
+    if (candidates.length === 0) {
+        return {
+            enabled: false,
+            skippedReason: "no-scip-supported-languages",
+            languages: [],
+            totalEdgesAdded: 0,
+            totalEdgesUpgraded: 0,
+            durationMs: Date.now() - start,
+        };
+    }
+    const outputDir = join(ctx.repoPath, META_DIR_NAME, "scip");
+    const offline = Boolean(ctx.options.offline);
+    const allowBuildScripts = process.env["CODEHUB_ALLOW_BUILD_SCRIPTS"] === "1";
+    const perLang = [];
+    const nodesByFile = indexNodesByFile(ctx);
+    // Fan out: per-language indexer runs are independent.
+    const results = await Promise.all(candidates.map(async (kind) => {
+        const scipPath = join(outputDir, `${kind}.scip`);
+        let result;
+        if (offline && !existsSync(scipPath)) {
+            result = {
+                kind,
+                scipPath,
+                tool: kindToTool(kind),
+                version: "",
+                skipped: true,
+                skipReason: "offline-no-cached-index",
+                durationMs: 0,
+            };
+        }
+        else if (isCacheFresh(scipPath, ctx.repoPath, kind)) {
+            result = {
+                kind,
+                scipPath,
+                tool: kindToTool(kind),
+                version: probeCachedVersion(scipPath),
+                skipped: false,
+                durationMs: 0,
+            };
+        }
+        else {
+            try {
+                result = await runIndexer(kind, {
+                    projectRoot: ctx.repoPath,
+                    outputDir,
+                    allowBuildScripts,
+                });
+            }
+            catch (err) {
+                ctx.onProgress?.({
+                    phase: SCIP_INDEX_PHASE_NAME,
+                    kind: "warn",
+                    message: `scip-index: ${kind} indexer failed — ${err.message}`,
+                });
+                result = {
+                    kind,
+                    scipPath,
+                    tool: kindToTool(kind),
+                    version: "",
+                    skipped: true,
+                    skipReason: `indexer-error:${err.message}`,
+                    durationMs: 0,
+                };
+            }
+        }
+        return result;
+    }));
+    let totalAdded = 0;
+    let totalUpgraded = 0;
+    const existingEdgeKeys = snapshotEdgeKeys(ctx);
+    for (const result of results) {
+        if (result.skipped) {
+            ctx.onProgress?.({
+                phase: SCIP_INDEX_PHASE_NAME,
+                kind: "warn",
+                message: `scip-index: ${result.kind} skipped — ${result.skipReason ?? "unknown"}`,
+            });
+            perLang.push({
+                kind: result.kind,
+                tool: result.tool,
+                version: result.version,
+                skipped: true,
+                skipReason: result.skipReason ?? "skipped",
+                scipPath: result.scipPath,
+                edgesAdded: 0,
+                edgesUpgraded: 0,
+            });
+            continue;
+        }
+        if (!existsSync(result.scipPath)) {
+            perLang.push({
+                kind: result.kind,
+                tool: result.tool,
+                version: result.version,
+                skipped: true,
+                skipReason: "scip-output-missing",
+                scipPath: result.scipPath,
+                edgesAdded: 0,
+                edgesUpgraded: 0,
+            });
+            continue;
+        }
+        const buf = await readFile(result.scipPath);
+        const index = parseScipIndex(new Uint8Array(buf));
+        const derived = deriveIndex(index);
+        const symbolDef = buildSymbolDefIndex(index);
+        const reason = scipProvenanceReason(kindToProvenance(result.kind), result.version || index.tool.version || "unknown");
+        const { added: edgeAdded, upgraded: edgeUpgraded } = emitEdges(ctx, nodesByFile, derived.edges, symbolDef, reason, existingEdgeKeys);
+        const { added: relAdded, upgraded: relUpgraded } = emitRelations(ctx, nodesByFile, derived.relations, symbolDef, reason, existingEdgeKeys);
+        const added = edgeAdded + relAdded;
+        const upgraded = edgeUpgraded + relUpgraded;
+        totalAdded += added;
+        totalUpgraded += upgraded;
+        perLang.push({
+            kind: result.kind,
+            tool: result.tool,
+            version: result.version || index.tool.version || "unknown",
+            skipped: false,
+            scipPath: result.scipPath,
+            edgesAdded: added,
+            edgesUpgraded: upgraded,
+        });
+    }
+    return {
+        enabled: perLang.some((p) => !p.skipped),
+        languages: perLang,
+        totalEdgesAdded: totalAdded,
+        totalEdgesUpgraded: totalUpgraded,
+        durationMs: Date.now() - start,
+    };
+}
+// ---- helpers ------------------------------------------------------------
+function findProfileNode(ctx) {
+    for (const n of ctx.graph.nodes()) {
+        if (n.kind === "ProjectProfile") {
+            const langs = n.languages ?? [];
+            return { languages: [...langs] };
+        }
+    }
+    return undefined;
+}
+function scipLangToOchLang(k) {
+    switch (k) {
+        case "typescript":
+            return "typescript";
+        case "python":
+            return "python";
+        case "go":
+            return "go";
+        case "rust":
+            return "rust";
+        case "java":
+            return "java";
+        case "clang":
+            // `clang` covers C + C++. Downstream LanguageId is a single token;
+            // "c" matches existing code paths that have looked up C-derived
+            // sources by extension. C++-specific consumers see `clang` under
+            // the indexer name in provenance reasons.
+            return "c";
+        case "cobol-proleap":
+            return "cobol";
+        case "ruby":
+            return "ruby";
+        case "dotnet":
+            return "csharp";
+        case "kotlin":
+            return "kotlin";
+    }
+}
+function kindToTool(k) {
+    return k === "rust" ? "rust-analyzer" : `scip-${k}`;
+}
+function kindToProvenance(k) {
+    switch (k) {
+        case "typescript":
+            return "scip-typescript";
+        case "python":
+            return "scip-python";
+        case "go":
+            return "scip-go";
+        case "rust":
+            return "rust-analyzer";
+        case "java":
+            return "scip-java";
+        case "clang":
+            return "scip-clang";
+        case "cobol-proleap":
+            // cobol-proleap edges don't flow through the SCIP derivation path —
+            // the in-process bridge emits CodeRelation rows directly. This switch
+            // arm exists only to keep the function exhaustive under
+            // `noFallthroughCasesInSwitch`; callers never invoke scipProvenance
+            // for the cobol-proleap kind.
+            return "scip-typescript";
+        case "ruby":
+            return "scip-ruby";
+        case "dotnet":
+            return "scip-dotnet";
+        case "kotlin":
+            return "scip-kotlin";
+    }
+}
+function isCacheFresh(scipPath, repoPath, _kind) {
+    if (!existsSync(scipPath))
+        return false;
+    // Coarse heuristic: if the .scip file exists and is newer than the
+    // repo root mtime, treat it as fresh. A finer walk would compare
+    // against every source file for the language, but that duplicates the
+    // scan phase work. Callers can force a rebuild by deleting the file
+    // or touching the repo root.
+    try {
+        const scipMtime = statSync(scipPath).mtimeMs;
+        const rootMtime = statSync(repoPath).mtimeMs;
+        return scipMtime >= rootMtime;
+    }
+    catch {
+        return false;
+    }
+}
+function probeCachedVersion(scipPath) {
+    try {
+        // We don't parse-scan twice — return "cached" and let the actual
+        // decode phase read the real tool version from Metadata. Callers
+        // replace this string when emitting edges.
+        return statSync(scipPath).mtimeMs ? "cached" : "";
+    }
+    catch {
+        return "";
+    }
+}
+const SCIP_SYMBOL_KINDS = new Set([
+    "Class",
+    "Method",
+    "Function",
+    "Interface",
+    "Struct",
+    "Enum",
+    "Trait",
+]);
+function indexNodesByFile(ctx) {
+    const map = new Map();
+    for (const n of ctx.graph.nodes()) {
+        if (!SCIP_SYMBOL_KINDS.has(n.kind))
+            continue;
+        const startLine = n.startLine;
+        const endLine = n.endLine;
+        if (startLine === undefined || endLine === undefined)
+            continue;
+        const rec = {
+            id: n.id,
+            filePath: n.filePath,
+            startLine,
+            endLine,
+        };
+        const arr = map.get(n.filePath);
+        if (arr === undefined)
+            map.set(n.filePath, [rec]);
+        else
+            arr.push(rec);
+    }
+    for (const arr of map.values()) {
+        arr.sort((a, b) => {
+            if (a.startLine !== b.startLine)
+                return a.startLine - b.startLine;
+            return a.endLine - b.endLine;
+        });
+    }
+    return map;
+}
+function findEnclosingNodeId(nodesByFile, filePath, line) {
+    const candidates = nodesByFile.get(filePath);
+    if (candidates === undefined)
+        return undefined;
+    let best;
+    let bestSpan = Number.POSITIVE_INFINITY;
+    for (const rec of candidates) {
+        if (rec.startLine > line)
+            break;
+        if (rec.endLine < line)
+            continue;
+        const span = rec.endLine - rec.startLine;
+        if (span < bestSpan) {
+            best = rec;
+            bestSpan = span;
+        }
+    }
+    return best?.id;
+}
+function edgeKey(from, type, to) {
+    return `${from}\x00${type}\x00${to}`;
+}
+function snapshotEdgeKeys(ctx) {
+    const s = new Set();
+    for (const e of ctx.graph.edges()) {
+        s.add(edgeKey(e.from, e.type, e.to));
+    }
+    return s;
+}
+function emitEdges(ctx, nodesByFile, edges, symbolDef, reason, existingKeys) {
+    let added = 0;
+    let upgraded = 0;
+    // SCIP symbol strings are not OpenCodeHub node ids. Every derived edge
+    // needs two lookups: the caller's enclosing OCH node at the call site
+    // `(e.document, e.callLine)`, and the callee's enclosing OCH node at
+    // the callee's actual definition site. `symbolDef` carries the
+    // definition `(file, line)` for every SCIP symbol that has a
+    // DEFINITION occurrence anywhere in the index, so callee resolution is
+    // disambiguated even when multiple in-repo symbols share a display
+    // name. Symbols without a DEFINITION occurrence are external
+    // (stdlib / vendored / absent typings) and their edges are dropped.
+    // Each edge carries its own `kind` (CALLS or REFERENCES) so this loop
+    // routes both function-call and read-side reference fanout through the
+    // same caller→callee join shape.
+    for (const e of edges) {
+        const fromId = findEnclosingNodeId(nodesByFile, e.document, e.callLine + 1);
+        if (!fromId)
+            continue;
+        const calleeDef = symbolDef.get(e.callee);
+        if (!calleeDef)
+            continue;
+        const toId = findEnclosingNodeId(nodesByFile, calleeDef.file, calleeDef.line + 1);
+        if (!toId)
+            continue;
+        if (fromId === toId)
+            continue;
+        const key = edgeKey(fromId, e.kind, toId);
+        const priorExists = existingKeys.has(key);
+        ctx.graph.addEdge({
+            from: fromId,
+            to: toId,
+            type: e.kind,
+            confidence: SCIP_CONFIDENCE,
+            reason,
+        });
+        existingKeys.add(key);
+        if (priorExists)
+            upgraded += 1;
+        else
+            added += 1;
+    }
+    return { added, upgraded };
+}
+/**
+ * Emit IMPLEMENTS / TYPE_OF graph edges from `derived.relations`.
+ *
+ * `collectRels` in `@opencodehub/scip-ingest/derive.ts` translates the
+ * SCIP `Relationship` message (`is_implementation`, `is_type_definition`)
+ * into structural relations between two SCIP symbols. Both ends need to
+ * resolve to OCH nodes via `symbolDef` — a relation whose source or
+ * target has no DEFINITION anywhere in the index is dropped (the
+ * relation lives entirely outside the indexed corpus). Otherwise the
+ * lookup uses the same `+1` boundary translation as `emitEdges` because
+ * SCIP `range.startLine` is 0-indexed and OCH graph nodes are 1-indexed.
+ */
+function emitRelations(ctx, nodesByFile, relations, symbolDef, reason, existingKeys) {
+    let added = 0;
+    let upgraded = 0;
+    for (const r of relations) {
+        const fromDef = symbolDef.get(r.from);
+        if (!fromDef)
+            continue;
+        const toDef = symbolDef.get(r.to);
+        if (!toDef)
+            continue;
+        const fromId = findEnclosingNodeId(nodesByFile, fromDef.file, fromDef.line + 1);
+        if (!fromId)
+            continue;
+        const toId = findEnclosingNodeId(nodesByFile, toDef.file, toDef.line + 1);
+        if (!toId)
+            continue;
+        if (fromId === toId)
+            continue;
+        const key = edgeKey(fromId, r.kind, toId);
+        const priorExists = existingKeys.has(key);
+        ctx.graph.addEdge({
+            from: fromId,
+            to: toId,
+            type: r.kind,
+            confidence: SCIP_CONFIDENCE,
+            reason,
+        });
+        existingKeys.add(key);
+        if (priorExists)
+            upgraded += 1;
+        else
+            added += 1;
+    }
+    return { added, upgraded };
+}
+//# sourceMappingURL=scip-index.js.map

package/dist/pipeline/phases/scip-index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"scip-index.js","sourceRoot":"","sources":["../../../src/pipeline/phases/scip-index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AAEH,OAAO,EAAE,UAAU,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAC;AAC/C,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAC5C,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AASjC,OAAO,EACL,mBAAmB,EACnB,WAAW,EACX,eAAe,EACf,cAAc,EACd,UAAU,EACV,oBAAoB,GACrB,MAAM,0BAA0B,CAAC;AAClC,OAAO,EAAE,aAAa,EAAE,MAAM,sBAAsB,CAAC;AAErD,OAAO,EAAE,mBAAmB,EAAE,MAAM,eAAe,CAAC;AACpD,OAAO,EAAE,qBAAqB,EAAE,MAAM,iBAAiB,CAAC;AACxD,OAAO,EAAE,gBAAgB,EAAE,MAAM,YAAY,CAAC;AAC9C,OAAO,EAAE,kBAAkB,EAAsB,MAAM,cAAc,CAAC;AACtE,OAAO,EAAE,eAAe,EAAE,MAAM,WAAW,CAAC;AAE5C,MAAM,CAAC,MAAM,qBAAqB,GAAG,YAAY,CAAC;AAElD,MAAM,eAAe,GAAG,GAAG,CAAC;AAsB5B,MAAM,CAAC,MAAM,cAAc,GAAmC;IAC5D,IAAI,EAAE,qBAAqB;IAC3B,IAAI,EAAE;QACJ,eAAe;QACf,kBAAkB;QAClB,gBAAgB;QAChB,qBAAqB;QACrB,mBAAmB;KACpB;IACD,KAAK,CAAC,GAAG,CAAC,GAAG,EAAE,IAAI;QACjB,OAAO,YAAY,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;IACjC,CAAC;CACF,CAAC;AAEF,KAAK,UAAU,YAAY,CACzB,GAAoB,EACpB,IAAkC;IAElC,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IAEzB,IAAI,OAAO,CAAC,GAAG,CAAC,sBAAsB,CAAC,KAAK,GAAG,EAAE,CAAC;QAChD,OAAO;YACL,OAAO,EAAE,KAAK;YACd,aAAa,EAAE,wBAAwB;YACvC,SAAS,EAAE,EAAE;YACb,eAAe,EAAE,CAAC;YAClB,kBAAkB,EAAE,CAAC;YACrB,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,KAAK;SAC/B,CAAC;IACJ,CAAC;IAED,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,kBAAkB,CAA8B,CAAC;IAC1E,MAAM,WAAW,GAAG,eAAe,CAAC,GAAG,CAAC,CAAC;IACzC,IAAI,OAAO,KAAK,SAAS,IAAI,WAAW,KAAK,SAAS,EAAE,CAAC;QACvD,OAAO;YACL,OAAO,EAAE,KAAK;YACd,aAAa,EAAE,mBAAmB;YAClC,SAAS,EAAE,EAAE;YACb,eAAe,EAAE,CAAC;YAClB,kBAAkB,EAAE,CAAC;YACrB,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,KAAK;SAC/B,CAAC;IACJ,CAAC;IAED,MAAM,gBAAgB,GAAG,IAAI,GAAG,CAAC,WAAW,CAAC,SAAS,CAAC,CAAC;IACxD,MAAM,UAAU,GAAG,eAAe,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,CAAC,CAAc,EAAE,EAAE,CACzE,gBAAgB,CAAC,GAAG,CAAC,iBAAiB,CAAC,CAAC,CAAC,CAAC,CAC3C,CAAC;IAEF,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC5B,OAAO;YACL,OAAO,EAAE,KAAK;YACd,aAAa,EAAE,6BAA6B;YAC5C,SAAS,EAAE,EAAE;YACb,eAAe,EAAE,CAAC;YAClB,kBAAkB,EAAE,CAAC;YACrB,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,KAAK;SAC/B,CAAC;IACJ,CAAC;IAED,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,QAAQ,EAAE,aAAa,EAAE,MAAM,CAAC,CAAC;IAC5D,MAAM,OAAO,GAAG,OAAO,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC;IAC7C,MAAM,iBAAiB,GAAG,OAAO,CAAC,GAAG,CAAC,6BAA6B,CAAC,KAAK,GAAG,CAAC;IAE7E,MAAM,OAAO,GAA2B,EAAE,CAAC;IAC3C,MAAM,WAAW,GAAG,gBAAgB,CAAC,GAAG,CAAC,CAAC;IAE1C,sDAAsD;IACtD,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,GAAG,CAC/B,UAAU,CAAC,GAAG,CAAC,KAAK,EAAE,IAAiB,EAAE,EAAE;QACzC,MAAM,QAAQ,GAAG,IAAI,CAAC,SAAS,EAAE,GAAG,IAAI,OAAO,CAAC,CAAC;QACjD,IAAI,MAAqB,CAAC;QAC1B,IAAI,OAAO,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;YACrC,MAAM,GAAG;gBACP,IAAI;gBACJ,QAAQ;gBACR,IAAI,EAAE,UAAU,CAAC,IAAI,CAAC;gBACtB,OAAO,EAAE,EAAE;gBACX,OAAO,EAAE,IAAI;gBACb,UAAU,EAAE,yBAAyB;gBACrC,UAAU,EAAE,CAAC;aACd,CAAC;QACJ,CAAC;aAAM,IAAI,YAAY,CAAC,QAAQ,EAAE,GAAG,CAAC,QAAQ,EAAE,IAAI,CAAC,EAAE,CAAC;YACtD,MAAM,GAAG;gBACP,IAAI;gBACJ,QAAQ;gBACR,IAAI,EAAE,UAAU,CAAC,IAAI,CAAC;gBACtB,OAAO,EAAE,kBAAkB,CAAC,QAAQ,CAAC;gBACrC,OAAO,EAAE,KAAK;gBACd,UAAU,EAAE,CAAC;aACd,CAAC;QACJ,CAAC;aAAM,CAAC;YACN,IAAI,CAAC;gBACH,MAAM,GAAG,MAAM,UAAU,CAAC,IAAI,EAAE;oBAC9B,WAAW,EAAE,GAAG,CAAC,QAAQ;oBACzB,SAAS;oBACT,iBAAiB;iBAClB,CAAC,CAAC;YACL,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,GAAG,CAAC,UAAU,EAAE,CAAC;oBACf,KAAK,EAAE,qBAAqB;oBAC5B,IAAI,EAAE,MAAM;oBACZ,OAAO,EAAE,eAAe,IAAI,qBAAsB,GAAa,CAAC,OAAO,EAAE;iBAC1E,CAAC,CAAC;gBACH,MAAM,GAAG;oBACP,IAAI;oBACJ,QAAQ;oBACR,IAAI,EAAE,UAAU,CAAC,IAAI,CAAC;oBACtB,OAAO,EAAE,EAAE;oBACX,OAAO,EAAE,IAAI;oBACb,UAAU,EAAE,iBAAkB,GAAa,CAAC,OAAO,EAAE;oBACrD,UAAU,EAAE,CAAC;iBACd,CAAC;YACJ,CAAC;QACH,CAAC;QACD,OAAO,MAAM,CAAC;IAChB,CAAC,CAAC,CACH,CAAC;IAEF,IAAI,UAAU,GAAG,CAAC,CAAC;IACnB,IAAI,aAAa,GAAG,CAAC,CAAC;IACtB,MAAM,gBAAgB,GAAG,gBAAgB,CAAC,GAAG,CAAC,CAAC;IAE/C,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,IAAI,MAAM,CAAC,OAAO,EAAE,CAAC;YACnB,GAAG,CAAC,UAAU,EAAE,CAAC;gBACf,KAAK,EAAE,qBAAqB;gBAC5B,IAAI,EAAE,MAAM;gBACZ,OAAO,EAAE,eAAe,MAAM,CAAC,IAAI,cAAc,MAAM,CAAC,UAAU,IAAI,SAAS,EAAE;aAClF,CAAC,CAAC;YACH,OAAO,CAAC,IAAI,CAAC;gBACX,IAAI,EAAE,MAAM,CAAC,IAAI;gBACjB,IAAI,EAAE,MAAM,CAAC,IAAI;gBACjB,OAAO,EAAE,MAAM,CAAC,OAAO;gBACvB,OAAO,EAAE,IAAI;gBACb,UAAU,EAAE,MAAM,CAAC,UAAU,IAAI,SAAS;gBAC1C,QAAQ,EAAE,MAAM,CAAC,QAAQ;gBACzB,UAAU,EAAE,CAAC;gBACb,aAAa,EAAE,CAAC;aACjB,CAAC,CAAC;YACH,SAAS;QACX,CAAC;QACD,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,CAAC;YACjC,OAAO,CAAC,IAAI,CAAC;gBACX,IAAI,EAAE,MAAM,CAAC,IAAI;gBACjB,IAAI,EAAE,MAAM,CAAC,IAAI;gBACjB,OAAO,EAAE,MAAM,CAAC,OAAO;gBACvB,OAAO,EAAE,IAAI;gBACb,UAAU,EAAE,qBAAqB;gBACjC,QAAQ,EAAE,MAAM,CAAC,QAAQ;gBACzB,UAAU,EAAE,CAAC;gBACb,aAAa,EAAE,CAAC;aACjB,CAAC,CAAC;YACH,SAAS;QACX,CAAC;QACD,MAAM,GAAG,GAAG,MAAM,QAAQ,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;QAC5C,MAAM,KAAK,GAAG,cAAc,CAAC,IAAI,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC;QAClD,MAAM,OAAO,GAAG,WAAW,CAAC,KAAK,CAAC,CAAC;QACnC,MAAM,SAAS,GAAG,mBAAmB,CAAC,KAAK,CAAC,CAAC;QAC7C,MAAM,MAAM,GAAG,oBAAoB,CACjC,gBAAgB,CAAC,MAAM,CAAC,IAAI,CAAC,EAC7B,MAAM,CAAC,OAAO,IAAI,KAAK,CAAC,IAAI,CAAC,OAAO,IAAI,SAAS,CAClD,CAAC;QAEF,MAAM,EAAE,KAAK,EAAE,SAAS,EAAE,QAAQ,EAAE,YAAY,EAAE,GAAG,SAAS,CAC5D,GAAG,EACH,WAAW,EACX,OAAO,CAAC,KAAK,EACb,SAAS,EACT,MAAM,EACN,gBAAgB,CACjB,CAAC;QACF,MAAM,EAAE,KAAK,EAAE,QAAQ,EAAE,QAAQ,EAAE,WAAW,EAAE,GAAG,aAAa,CAC9D,GAAG,EACH,WAAW,EACX,OAAO,CAAC,SAAS,EACjB,SAAS,EACT,MAAM,EACN,gBAAgB,CACjB,CAAC;QACF,MAAM,KAAK,GAAG,SAAS,GAAG,QAAQ,CAAC;QACnC,MAAM,QAAQ,GAAG,YAAY,GAAG,WAAW,CAAC;QAC5C,UAAU,IAAI,KAAK,CAAC;QACpB,aAAa,IAAI,QAAQ,CAAC;QAC1B,OAAO,CAAC,IAAI,CAAC;YACX,IAAI,EAAE,MAAM,CAAC,IAAI;YACjB,IAAI,EAAE,MAAM,CAAC,IAAI;YACjB,OAAO,EAAE,MAAM,CAAC,OAAO,IAAI,KAAK,CAAC,IAAI,CAAC,OAAO,IAAI,SAAS;YAC1D,OAAO,EAAE,KAAK;YACd,QAAQ,EAAE,MAAM,CAAC,QAAQ;YACzB,UAAU,EAAE,KAAK;YACjB,aAAa,EAAE,QAAQ;SACxB,CAAC,CAAC;IACL,CAAC;IAED,OAAO;QACL,OAAO,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC;QACxC,SAAS,EAAE,OAAO;QAClB,eAAe,EAAE,UAAU;QAC3B,kBAAkB,EAAE,aAAa;QACjC,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,KAAK;KAC/B,CAAC;AACJ,CAAC;AAED,4EAA4E;AAE5E,SAAS,eAAe,CAAC,GAAoB;IAC3C,KAAK,MAAM,CAAC,IAAI,GAAG,CAAC,KAAK,CAAC,KAAK,EAAE,EAAE,CAAC;QAClC,IAAI,CAAC,CAAC,IAAI,KAAK,gBAAgB,EAAE,CAAC;YAChC,MAAM,KAAK,GAAI,CAAuC,CAAC,SAAS,IAAI,EAAE,CAAC;YACvE,OAAO,EAAE,SAAS,EAAE,CAAC,GAAG,KAAK,CAAC,EAAE,CAAC;QACnC,CAAC;IACH,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC;AAMD,SAAS,iBAAiB,CAAC,CAAc;IACvC,QAAQ,CAAC,EAAE,CAAC;QACV,KAAK,YAAY;YACf,OAAO,YAAY,CAAC;QACtB,KAAK,QAAQ;YACX,OAAO,QAAQ,CAAC;QAClB,KAAK,IAAI;YACP,OAAO,IAAI,CAAC;QACd,KAAK,MAAM;YACT,OAAO,MAAM,CAAC;QAChB,KAAK,MAAM;YACT,OAAO,MAAM,CAAC;QAChB,KAAK,OAAO;YACV,mEAAmE;YACnE,gEAAgE;YAChE,iEAAiE;YACjE,0CAA0C;YAC1C,OAAO,GAAG,CAAC;QACb,KAAK,eAAe;YAClB,OAAO,OAAO,CAAC;QACjB,KAAK,MAAM;YACT,OAAO,MAAM,CAAC;QAChB,KAAK,QAAQ;YACX,OAAO,QAAQ,CAAC;QAClB,KAAK,QAAQ;YACX,OAAO,QAAQ,CAAC;IACpB,CAAC;AACH,CAAC;AAED,SAAS,UAAU,CAAC,CAAc;IAChC,OAAO,CAAC,KAAK,MAAM,CAAC,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,QAAQ,CAAC,EAAE,CAAC;AACtD,CAAC;AAED,SAAS,gBAAgB,CAAC,CAAc;IACtC,QAAQ,CAAC,EAAE,CAAC;QACV,KAAK,YAAY;YACf,OAAO,iBAAiB,CAAC;QAC3B,KAAK,QAAQ;YACX,OAAO,aAAa,CAAC;QACvB,KAAK,IAAI;YACP,OAAO,SAAS,CAAC;QACnB,KAAK,MAAM;YACT,OAAO,eAAe,CAAC;QACzB,KAAK,MAAM;YACT,OAAO,WAAW,CAAC;QACrB,KAAK,OAAO;YACV,OAAO,YAAY,CAAC;QACtB,KAAK,eAAe;YAClB,oEAAoE;YACpE,sEAAsE;YACtE,wDAAwD;YACxD,oEAAoE;YACpE,8BAA8B;YAC9B,OAAO,iBAAiB,CAAC;QAC3B,KAAK,MAAM;YACT,OAAO,WAAW,CAAC;QACrB,KAAK,QAAQ;YACX,OAAO,aAAa,CAAC;QACvB,KAAK,QAAQ;YACX,OAAO,aAAa,CAAC;IACzB,CAAC;AACH,CAAC;AAED,SAAS,YAAY,CAAC,QAAgB,EAAE,QAAgB,EAAE,KAAkB;IAC1E,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC;QAAE,OAAO,KAAK,CAAC;IACxC,mEAAmE;IACnE,iEAAiE;IACjE,sEAAsE;IACtE,oEAAoE;IACpE,6BAA6B;IAC7B,IAAI,CAAC;QACH,MAAM,SAAS,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAC,OAAO,CAAC;QAC7C,MAAM,SAAS,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAC,OAAO,CAAC;QAC7C,OAAO,SAAS,IAAI,SAAS,CAAC;IAChC,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,KAAK,CAAC;IACf,CAAC;AACH,CAAC;AAED,SAAS,kBAAkB,CAAC,QAAgB;IAC1C,IAAI,CAAC;QACH,iEAAiE;QACjE,iEAAiE;QACjE,2CAA2C;QAC3C,OAAO,QAAQ,CAAC,QAAQ,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,CAAC;IACpD,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,EAAE,CAAC;IACZ,CAAC;AACH,CAAC;AAWD,MAAM,iBAAiB,GAAwB,IAAI,GAAG,CAAC;IACrD,OAAO;IACP,QAAQ;IACR,UAAU;IACV,WAAW;IACX,QAAQ;IACR,MAAM;IACN,OAAO;CACR,CAAC,CAAC;AAEH,SAAS,gBAAgB,CAAC,GAAoB;IAC5C,MAAM,GAAG,GAAG,IAAI,GAAG,EAAuB,CAAC;IAC3C,KAAK,MAAM,CAAC,IAAI,GAAG,CAAC,KAAK,CAAC,KAAK,EAAE,EAAE,CAAC;QAClC,IAAI,CAAC,iBAAiB,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC;YAAE,SAAS;QAC7C,MAAM,SAAS,GAAI,CAA4B,CAAC,SAAS,CAAC;QAC1D,MAAM,OAAO,GAAI,CAA0B,CAAC,OAAO,CAAC;QACpD,IAAI,SAAS,KAAK,SAAS,IAAI,OAAO,KAAK,SAAS;YAAE,SAAS;QAC/D,MAAM,GAAG,GAAc;YACrB,EAAE,EAAE,CAAC,CAAC,EAAY;YAClB,QAAQ,EAAE,CAAC,CAAC,QAAQ;YACpB,SAAS;YACT,OAAO;SACR,CAAC;QACF,MAAM,GAAG,GAAG,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC;QAChC,IAAI,GAAG,KAAK,SAAS;YAAE,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,QAAQ,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;;YAC7C,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACrB,CAAC;IACD,KAAK,MAAM,GAAG,IAAI,GAAG,CAAC,MAAM,EAAE,EAAE,CAAC;QAC/B,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;YAChB,IAAI,CAAC,CAAC,SAAS,KAAK,CAAC,CAAC,SAAS;gBAAE,OAAO,CAAC,CAAC,SAAS,GAAG,CAAC,CAAC,SAAS,CAAC;YAClE,OAAO,CAAC,CAAC,OAAO,GAAG,CAAC,CAAC,OAAO,CAAC;QAC/B,CAAC,CAAC,CAAC;IACL,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,mBAAmB,CAC1B,WAAwB,EACxB,QAAgB,EAChB,IAAY;IAEZ,MAAM,UAAU,GAAG,WAAW,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;IAC7C,IAAI,UAAU,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC/C,IAAI,IAA2B,CAAC;IAChC,IAAI,QAAQ,GAAG,MAAM,CAAC,iBAAiB,CAAC;IACxC,KAAK,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;QAC7B,IAAI,GAAG,CAAC,SAAS,GAAG,IAAI;YAAE,MAAM;QAChC,IAAI,GAAG,CAAC,OAAO,GAAG,IAAI;YAAE,SAAS;QACjC,MAAM,IAAI,GAAG,GAAG,CAAC,OAAO,GAAG,GAAG,CAAC,SAAS,CAAC;QACzC,IAAI,IAAI,GAAG,QAAQ,EAAE,CAAC;YACpB,IAAI,GAAG,GAAG,CAAC;YACX,QAAQ,GAAG,IAAI,CAAC;QAClB,CAAC;IACH,CAAC;IACD,OAAO,IAAI,EAAE,EAAE,CAAC;AAClB,CAAC;AAED,SAAS,OAAO,CAAC,IAAY,EAAE,IAAY,EAAE,EAAU;IACrD,OAAO,GAAG,IAAI,OAAO,IAAI,OAAO,EAAE,EAAE,CAAC;AACvC,CAAC;AAED,SAAS,gBAAgB,CAAC,GAAoB;IAC5C,MAAM,CAAC,GAAG,IAAI,GAAG,EAAU,CAAC;IAC5B,KAAK,MAAM,CAAC,IAAI,GAAG,CAAC,KAAK,CAAC,KAAK,EAAE,EAAE,CAAC;QAClC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,IAAc,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,EAAY,CAAC,CAAC,CAAC;IAC3D,CAAC;IACD,OAAO,CAAC,CAAC;AACX,CAAC;AAED,SAAS,SAAS,CAChB,GAAoB,EACpB,WAAwB,EACxB,KAA6B,EAC7B,SAA8D,EAC9D,MAAc,EACd,YAAyB;IAEzB,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,QAAQ,GAAG,CAAC,CAAC;IACjB,uEAAuE;IACvE,sEAAsE;IACtE,qEAAqE;IACrE,+DAA+D;IAC/D,6DAA6D;IAC7D,uEAAuE;IACvE,mEAAmE;IACnE,6DAA6D;IAC7D,oEAAoE;IACpE,sEAAsE;IACtE,uEAAuE;IACvE,iCAAiC;IACjC,KAAK,MAAM,CAAC,IAAI,KAAK,EAAE,CAAC;QACtB,MAAM,MAAM,GAAG,mBAAmB,CAAC,WAAW,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,CAAC,QAAQ,GAAG,CAAC,CAAC,CAAC;QAC5E,IAAI,CAAC,MAAM;YAAE,SAAS;QACtB,MAAM,SAAS,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;QAC1C,IAAI,CAAC,SAAS;YAAE,SAAS;QACzB,MAAM,IAAI,GAAG,mBAAmB,CAAC,WAAW,EAAE,SAAS,CAAC,IAAI,EAAE,SAAS,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC;QAClF,IAAI,CAAC,IAAI;YAAE,SAAS;QACpB,IAAI,MAAM,KAAK,IAAI;YAAE,SAAS;QAE9B,MAAM,GAAG,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;QAC1C,MAAM,WAAW,GAAG,YAAY,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QAE1C,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC;YAChB,IAAI,EAAE,MAAM;YACZ,EAAE,EAAE,IAAI;YACR,IAAI,EAAE,CAAC,CAAC,IAAI;YACZ,UAAU,EAAE,eAAe;YAC3B,MAAM;SACP,CAAC,CAAC;QAEH,YAAY,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QACtB,IAAI,WAAW;YAAE,QAAQ,IAAI,CAAC,CAAC;;YAC1B,KAAK,IAAI,CAAC,CAAC;IAClB,CAAC;IACD,OAAO,EAAE,KAAK,EAAE,QAAQ,EAAE,CAAC;AAC7B,CAAC;AAED;;;;;;;;;;;GAWG;AACH,SAAS,aAAa,CACpB,GAAoB,EACpB,WAAwB,EACxB,SAAqC,EACrC,SAA8D,EAC9D,MAAc,EACd,YAAyB;IAEzB,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,QAAQ,GAAG,CAAC,CAAC;IACjB,KAAK,MAAM,CAAC,IAAI,SAAS,EAAE,CAAC;QAC1B,MAAM,OAAO,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QACtC,IAAI,CAAC,OAAO;YAAE,SAAS;QACvB,MAAM,KAAK,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;QAClC,IAAI,CAAC,KAAK;YAAE,SAAS;QACrB,MAAM,MAAM,GAAG,mBAAmB,CAAC,WAAW,EAAE,OAAO,CAAC,IAAI,EAAE,OAAO,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC;QAChF,IAAI,CAAC,MAAM;YAAE,SAAS;QACtB,MAAM,IAAI,GAAG,mBAAmB,CAAC,WAAW,EAAE,KAAK,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC;QAC1E,IAAI,CAAC,IAAI;YAAE,SAAS;QACpB,IAAI,MAAM,KAAK,IAAI;YAAE,SAAS;QAE9B,MAAM,GAAG,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;QAC1C,MAAM,WAAW,GAAG,YAAY,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QAE1C,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC;YAChB,IAAI,EAAE,MAAM;YACZ,EAAE,EAAE,IAAI;YACR,IAAI,EAAE,CAAC,CAAC,IAAI;YACZ,UAAU,EAAE,eAAe;YAC3B,MAAM;SACP,CAAC,CAAC;QAEH,YAAY,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QACtB,IAAI,WAAW;YAAE,QAAQ,IAAI,CAAC,CAAC;;YAC1B,KAAK,IAAI,CAAC,CAAC;IAClB,CAAC;IACD,OAAO,EAAE,KAAK,EAAE,QAAQ,EAAE,CAAC;AAC7B,CAAC"}

package/dist/pipeline/phases/structure.d.ts

@@ -0,0 +1,21 @@
+/**
+ * Structure phase — projects the filesystem spine onto the graph.
+ *
+ * Consumes the scan phase output and emits:
+ *  - One `File` node per scanned file (id includes the relative path so
+ *    downstream IDs compose deterministically).
+ *  - One `Folder` node per unique ancestor directory.
+ *  - `CONTAINS` edges from each parent → child along the folder tree.
+ *
+ * Determinism: relative paths are sorted before processing so node and
+ * edge insertion order is reproducible across machines.
+ */
+import type { PipelinePhase } from "../types.js";
+export interface StructureOutput {
+    readonly pathSet: ReadonlySet<string>;
+    readonly folderCount: number;
+    readonly fileCount: number;
+}
+export declare const STRUCTURE_PHASE_NAME = "structure";
+export declare const structurePhase: PipelinePhase<StructureOutput>;
+//# sourceMappingURL=structure.d.ts.map

package/dist/pipeline/phases/structure.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"structure.d.ts","sourceRoot":"","sources":["../../../src/pipeline/phases/structure.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAIH,OAAO,KAAK,EAAmB,aAAa,EAAE,MAAM,aAAa,CAAC;AAGlE,MAAM,WAAW,eAAe;IAC9B,QAAQ,CAAC,OAAO,EAAE,WAAW,CAAC,MAAM,CAAC,CAAC;IACtC,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAC;IAC7B,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;CAC5B;AAED,eAAO,MAAM,oBAAoB,cAAc,CAAC;AAEhD,eAAO,MAAM,cAAc,EAAE,aAAa,CAAC,eAAe,CAUzD,CAAC"}

package/dist/pipeline/phases/structure.js

@@ -0,0 +1,115 @@
+/**
+ * Structure phase — projects the filesystem spine onto the graph.
+ *
+ * Consumes the scan phase output and emits:
+ *  - One `File` node per scanned file (id includes the relative path so
+ *    downstream IDs compose deterministically).
+ *  - One `Folder` node per unique ancestor directory.
+ *  - `CONTAINS` edges from each parent → child along the folder tree.
+ *
+ * Determinism: relative paths are sorted before processing so node and
+ * edge insertion order is reproducible across machines.
+ */
+import { makeNodeId } from "@opencodehub/core-types";
+import { SCAN_PHASE_NAME } from "./scan.js";
+export const STRUCTURE_PHASE_NAME = "structure";
+export const structurePhase = {
+    name: STRUCTURE_PHASE_NAME,
+    deps: [SCAN_PHASE_NAME],
+    async run(ctx, deps) {
+        const scan = deps.get(SCAN_PHASE_NAME);
+        if (scan === undefined) {
+            throw new Error("structure: scan output missing from dependency map");
+        }
+        return runStructure(ctx, scan);
+    },
+};
+function runStructure(ctx, scan) {
+    const relPaths = [...scan.files.map((f) => f.relPath)].sort();
+    // Pre-compute the set of every folder path that must exist.
+    const folderPaths = new Set();
+    for (const rel of relPaths) {
+        let parent = parentDir(rel);
+        while (parent !== "") {
+            folderPaths.add(parent);
+            parent = parentDir(parent);
+        }
+        // Root folder handled as an implicit anchor; we expose it as `"."` so
+        // it has a stable qualified name without being confused with files.
+    }
+    const sortedFolders = [...folderPaths].sort();
+    const rootFolder = ".";
+    folderPaths.add(rootFolder);
+    // Emit the root folder first so later CONTAINS edges point at an
+    // existing node.
+    const rootNode = {
+        id: makeNodeId("Folder", rootFolder, rootFolder),
+        kind: "Folder",
+        name: rootFolder,
+        filePath: rootFolder,
+    };
+    ctx.graph.addNode(rootNode);
+    for (const folder of sortedFolders) {
+        const node = {
+            id: makeNodeId("Folder", folder, folder),
+            kind: "Folder",
+            name: basename(folder),
+            filePath: folder,
+        };
+        ctx.graph.addNode(node);
+    }
+    // CONTAINS edges from parent folder to child folder.
+    for (const folder of sortedFolders) {
+        const parent = parentDir(folder);
+        const parentPath = parent === "" ? rootFolder : parent;
+        ctx.graph.addEdge({
+            from: makeNodeId("Folder", parentPath, parentPath),
+            to: makeNodeId("Folder", folder, folder),
+            type: "CONTAINS",
+            confidence: 1,
+            reason: "folder-to-folder",
+        });
+    }
+    // File nodes + CONTAINS from owning folder.
+    const pathSet = new Set();
+    for (const file of scan.files) {
+        pathSet.add(file.relPath);
+        const fileId = makeNodeId("File", file.relPath, file.relPath);
+        const fileNode = {
+            id: fileId,
+            kind: "File",
+            name: basename(file.relPath),
+            filePath: file.relPath,
+            contentHash: file.sha256,
+            ...(file.language !== undefined ? { language: file.language } : {}),
+        };
+        ctx.graph.addNode(fileNode);
+        const parent = parentDir(file.relPath);
+        const parentPath = parent === "" ? rootFolder : parent;
+        ctx.graph.addEdge({
+            from: makeNodeId("Folder", parentPath, parentPath),
+            to: fileId,
+            type: "CONTAINS",
+            confidence: 1,
+            reason: "folder-to-file",
+        });
+    }
+    return {
+        pathSet,
+        folderCount: folderPaths.size,
+        fileCount: scan.files.length,
+    };
+}
+function parentDir(p) {
+    const idx = p.lastIndexOf("/");
+    if (idx <= 0)
+        return "";
+    return p.slice(0, idx);
+}
+function basename(p) {
+    const idx = p.lastIndexOf("/");
+    if (idx < 0)
+        return p;
+    return p.slice(idx + 1);
+}
+//# sourceMappingURL=structure.js.map

package/dist/pipeline/phases/structure.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"structure.js","sourceRoot":"","sources":["../../../src/pipeline/phases/structure.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAGH,OAAO,EAAE,UAAU,EAAE,MAAM,yBAAyB,CAAC;AAErD,OAAO,EAAE,eAAe,EAAmB,MAAM,WAAW,CAAC;AAQ7D,MAAM,CAAC,MAAM,oBAAoB,GAAG,WAAW,CAAC;AAEhD,MAAM,CAAC,MAAM,cAAc,GAAmC;IAC5D,IAAI,EAAE,oBAAoB;IAC1B,IAAI,EAAE,CAAC,eAAe,CAAC;IACvB,KAAK,CAAC,GAAG,CAAC,GAAG,EAAE,IAAI;QACjB,MAAM,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,eAAe,CAA2B,CAAC;QACjE,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;YACvB,MAAM,IAAI,KAAK,CAAC,oDAAoD,CAAC,CAAC;QACxE,CAAC;QACD,OAAO,YAAY,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;IACjC,CAAC;CACF,CAAC;AAEF,SAAS,YAAY,CAAC,GAAoB,EAAE,IAAgB;IAC1D,MAAM,QAAQ,GAAG,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IAE9D,4DAA4D;IAC5D,MAAM,WAAW,GAAG,IAAI,GAAG,EAAU,CAAC;IACtC,KAAK,MAAM,GAAG,IAAI,QAAQ,EAAE,CAAC;QAC3B,IAAI,MAAM,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC;QAC5B,OAAO,MAAM,KAAK,EAAE,EAAE,CAAC;YACrB,WAAW,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;YACxB,MAAM,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC;QAC7B,CAAC;QACD,sEAAsE;QACtE,oEAAoE;IACtE,CAAC;IAED,MAAM,aAAa,GAAG,CAAC,GAAG,WAAW,CAAC,CAAC,IAAI,EAAE,CAAC;IAC9C,MAAM,UAAU,GAAG,GAAG,CAAC;IACvB,WAAW,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;IAE5B,iEAAiE;IACjE,iBAAiB;IACjB,MAAM,QAAQ,GAAe;QAC3B,EAAE,EAAE,UAAU,CAAC,QAAQ,EAAE,UAAU,EAAE,UAAU,CAAC;QAChD,IAAI,EAAE,QAAQ;QACd,IAAI,EAAE,UAAU;QAChB,QAAQ,EAAE,UAAU;KACrB,CAAC;IACF,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC;IAE5B,KAAK,MAAM,MAAM,IAAI,aAAa,EAAE,CAAC;QACnC,MAAM,IAAI,GAAe;YACvB,EAAE,EAAE,UAAU,CAAC,QAAQ,EAAE,MAAM,EAAE,MAAM,CAAC;YACxC,IAAI,EAAE,QAAQ;YACd,IAAI,EAAE,QAAQ,CAAC,MAAM,CAAC;YACtB,QAAQ,EAAE,MAAM;SACjB,CAAC;QACF,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;IAC1B,CAAC;IAED,qDAAqD;IACrD,KAAK,MAAM,MAAM,IAAI,aAAa,EAAE,CAAC;QACnC,MAAM,MAAM,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC;QACjC,MAAM,UAAU,GAAG,MAAM,KAAK,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,MAAM,CAAC;QACvD,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC;YAChB,IAAI,EAAE,UAAU,CAAC,QAAQ,EAAE,UAAU,EAAE,UAAU,CAAC;YAClD,EAAE,EAAE,UAAU,CAAC,QAAQ,EAAE,MAAM,EAAE,MAAM,CAAC;YACxC,IAAI,EAAE,UAAU;YAChB,UAAU,EAAE,CAAC;YACb,MAAM,EAAE,kBAAkB;SAC3B,CAAC,CAAC;IACL,CAAC;IAED,4CAA4C;IAC5C,MAAM,OAAO,GAAG,IAAI,GAAG,EAAU,CAAC;IAClC,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;QAC9B,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QAC1B,MAAM,MAAM,GAAG,UAAU,CAAC,MAAM,EAAE,IAAI,CAAC,OAAO,EAAE,IAAI,CAAC,OAAO,CAAC,CAAC;QAC9D,MAAM,QAAQ,GAAa;YACzB,EAAE,EAAE,MAAM;YACV,IAAI,EAAE,MAAM;YACZ,IAAI,EAAE,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC;YAC5B,QAAQ,EAAE,IAAI,CAAC,OAAO;YACtB,WAAW,EAAE,IAAI,CAAC,MAAM;YACxB,GAAG,CAAC,IAAI,CAAC,QAAQ,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,IAAI,CAAC,QAAQ,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;SACpE,CAAC;QACF,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC;QAE5B,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QACvC,MAAM,UAAU,GAAG,MAAM,KAAK,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,MAAM,CAAC;QACvD,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC;YAChB,IAAI,EAAE,UAAU,CAAC,QAAQ,EAAE,UAAU,EAAE,UAAU,CAAC;YAClD,EAAE,EAAE,MAAM;YACV,IAAI,EAAE,UAAU;YAChB,UAAU,EAAE,CAAC;YACb,MAAM,EAAE,gBAAgB;SACzB,CAAC,CAAC;IACL,CAAC;IAED,OAAO;QACL,OAAO;QACP,WAAW,EAAE,WAAW,CAAC,IAAI;QAC7B,SAAS,EAAE,IAAI,CAAC,KAAK,CAAC,MAAM;KAC7B,CAAC;AACJ,CAAC;AAED,SAAS,SAAS,CAAC,CAAS;IAC1B,MAAM,GAAG,GAAG,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;IAC/B,IAAI,GAAG,IAAI,CAAC;QAAE,OAAO,EAAE,CAAC;IACxB,OAAO,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;AACzB,CAAC;AAED,SAAS,QAAQ,CAAC,CAAS;IACzB,MAAM,GAAG,GAAG,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;IAC/B,IAAI,GAAG,GAAG,CAAC;QAAE,OAAO,CAAC,CAAC;IACtB,OAAO,CAAC,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC;AAC1B,CAAC"}