mind-palace-graph 0.3.0

package/dist/sources.js

@@ -0,0 +1,457 @@
+/**
+ * Source resolution.
+ *
+ * A "source" is a stream of text we can search. mpg supports four kinds:
+ *
+ *   - file/glob: read from disk
+ *   - command:   exec a shell command, search its stdout
+ *   - stdin:     read piped input
+ *   - url:       fetch with HTTP GET
+ *
+ * For non-file sources we capture the content into memory and feed it
+ * to rg via a temp file (see rg.ts). This keeps rg as the single
+ * search engine while supporting arbitrary content types.
+ */
+import { spawn } from "node:child_process";
+import { existsSync, statSync } from "node:fs";
+import { resolve as resolvePath } from "node:path";
+/** Cached stdin content so @- and content stdin don't double-read. */
+let _cachedStdin = null;
+/** Read stdin once and cache it. Returns cached value on subsequent calls. */
+export async function getStdin() {
+    if (_cachedStdin !== null)
+        return _cachedStdin;
+    if (process.stdin.isTTY) {
+        _cachedStdin = "";
+        return "";
+    }
+    const chunks = [];
+    for await (const chunk of process.stdin) {
+        chunks.push(typeof chunk === "string" ? Buffer.from(chunk) : chunk);
+    }
+    _cachedStdin = Buffer.concat(chunks).toString("utf8");
+    return _cachedStdin;
+}
+/** Reset the cached stdin (e.g. in test teardown). */
+export function resetStdinCache() {
+    _cachedStdin = null;
+}
+/** Expand globs into individual file paths using Node's built-in fs.glob. */
+export async function expandGlobs(patterns) {
+    if (patterns.length === 0)
+        return [];
+    const { glob } = await import("node:fs/promises");
+    const { statSync } = await import("node:fs");
+    const out = new Set();
+    async function walk(p) {
+        // If p is a directory, recurse with a trailing /** pattern.
+        // If p is a file, add it.
+        // If p contains wildcards, glob it.
+        let s = null;
+        try {
+            s = statSync(p);
+        }
+        catch { /* not on disk */ }
+        // Node's fs.glob treats `\` as a glob escape, so absolute Windows
+        // paths with backslashes never match. Normalize separators for the
+        // pattern only — entries returned by glob are still valid paths.
+        const toGlobPattern = (s) => s.replace(/\\/g, "/");
+        if (s && s.isDirectory()) {
+            const trimmed = p.replace(/[\\\/]+$/, "");
+            const recursePattern = `${toGlobPattern(trimmed)}/**`;
+            try {
+                for await (const entry of glob(recursePattern)) {
+                    try {
+                        const es = statSync(entry);
+                        if (es.isFile())
+                            out.add(entry);
+                    }
+                    catch { /* skip */ }
+                }
+            }
+            catch { /* ignore */ }
+        }
+        else if (s && s.isFile()) {
+            out.add(p);
+        }
+        else {
+            // Not on disk — treat as a glob pattern.
+            try {
+                for await (const entry of glob(toGlobPattern(p))) {
+                    try {
+                        const es = statSync(entry);
+                        if (es.isFile())
+                            out.add(entry);
+                    }
+                    catch { /* skip */ }
+                }
+            }
+            catch { /* ignore */ }
+        }
+    }
+    for (const pattern of patterns) {
+        await walk(pattern);
+    }
+    return [...out];
+}
+/** Heuristic: if a path exists as a file/dir, classify it. */
+export function classifyPath(p) {
+    if (!existsSync(p))
+        return "glob";
+    const s = statSync(p);
+    if (s.isDirectory())
+        return "glob";
+    if (s.isFile())
+        return "file";
+    return "glob";
+}
+/**
+ * Resolve a list of path specs to actual file paths.
+ *
+ * If `stdinContent` is provided, it is used as the content when
+ * resolving `@-` specs instead of reading process.stdin again.
+ * This avoids double-reading stdin when both content-from-stdin and
+ * path-list-from-stdin are used in the same invocation.
+ *
+ * A spec can be:
+ *   - `@-`         read paths from stdin, one per line
+ *   - `@<file>`    read paths from a file, one per line
+ *   - `path`       a literal file or directory path
+ *   - `glob`       a glob pattern; expanded via fs.glob
+ *
+ * Directories are recursed into. Empty lines and `#` comments are
+ * ignored when reading from a file.
+ */
+export async function resolvePathSpecs(specs, stdinContent) {
+    const out = new Set();
+    for (const spec of specs) {
+        if (spec === "@-") {
+            // Read paths from stdin.
+            const text = stdinContent ?? await getStdin();
+            for (const line of text.split(/\r?\n/)) {
+                const trimmed = line.trim();
+                if (!trimmed || trimmed.startsWith("#"))
+                    continue;
+                await addExpanded(trimmed, out);
+            }
+            continue;
+        }
+        if (spec.startsWith("@")) {
+            // Read paths from a file.
+            const filePath = spec.slice(1);
+            let text;
+            try {
+                const { readFileSync } = await import("node:fs");
+                text = readFileSync(filePath, "utf8");
+            }
+            catch (err) {
+                throw new Error(`Cannot read path list from @${filePath}: ${err.message}`);
+            }
+            for (const line of text.split(/\r?\n/)) {
+                const trimmed = line.trim();
+                if (!trimmed || trimmed.startsWith("#"))
+                    continue;
+                await addExpanded(trimmed, out);
+            }
+            continue;
+        }
+        await addExpanded(spec, out);
+    }
+    return [...out];
+}
+/**
+ * Split path specs into two buckets without expanding directories:
+ *
+ *   - `files`:    literal file paths the caller asked for. Each becomes
+ *                 a separate `runRg` invocation so they can be searched
+ *                 in parallel and their per-file content cache is hot.
+ *   - `bulk`:     directories and glob patterns. These get passed to
+ *                 rg as-is — rg walks them itself in parallel, much
+ *                 faster than fan-out-per-file from Node. Each bulk
+ *                 entry becomes one `runRg` invocation that may emit
+ *                 matches from many files.
+ *
+ * `@file` / `@-` are still expanded inline (the caller asked for an
+ * explicit list, so we respect that).
+ *
+ * Returns absolute paths so deduplication is stable across cwd-relative
+ * vs absolute inputs.
+ */
+export async function classifyPathSpecs(specs, stdinContent) {
+    const files = new Set();
+    const bulk = new Set();
+    function hasGlobMeta(s) {
+        // Match characters that imply globbing. We don't try to handle
+        // brace expansion (`{a,b}`) — rg doesn't accept it on argv either,
+        // so we expand ourselves below.
+        return /[*?\[\]]/.test(s);
+    }
+    async function classify(spec) {
+        if (existsSync(spec)) {
+            const s = statSync(spec);
+            if (s.isFile()) {
+                files.add(resolvePath(spec));
+                return;
+            }
+            if (s.isDirectory()) {
+                // The big win: directories go to rg as-is. rg walks them
+                // itself in parallel, much faster than one rg invocation per
+                // file from Node.
+                bulk.add(resolvePath(spec));
+                return;
+            }
+        }
+        if (hasGlobMeta(spec)) {
+            // rg does NOT accept shell-style globs as path args. Expand
+            // ourselves to literal files so the search target list stays
+            // valid. The win against the old code is that *dirs* (the
+            // common case) now skip expansion.
+            const expanded = await expandGlobs([spec]);
+            for (const f of expanded)
+                files.add(resolvePath(f));
+            return;
+        }
+        // Spec doesn't exist and has no glob meta — most likely a typo or
+        // a stale stash entry. Let rg surface the error rather than
+        // swallowing it silently.
+        bulk.add(spec);
+    }
+    for (const spec of specs) {
+        if (spec === "@-") {
+            const text = stdinContent ?? await getStdin();
+            for (const line of text.split(/\r?\n/)) {
+                const trimmed = line.trim();
+                if (!trimmed || trimmed.startsWith("#"))
+                    continue;
+                await classify(trimmed);
+            }
+            continue;
+        }
+        if (spec.startsWith("@")) {
+            const filePath = spec.slice(1);
+            let text;
+            try {
+                const { readFileSync } = await import("node:fs");
+                text = readFileSync(filePath, "utf8");
+            }
+            catch (err) {
+                throw new Error(`Cannot read path list from @${filePath}: ${err.message}`);
+            }
+            for (const line of text.split(/\r?\n/)) {
+                const trimmed = line.trim();
+                if (!trimmed || trimmed.startsWith("#"))
+                    continue;
+                await classify(trimmed);
+            }
+            continue;
+        }
+        await classify(spec);
+    }
+    return { files: [...files], bulk: [...bulk] };
+}
+/** Classify a spec, expand globs, and add all files to `out`. */
+async function addExpanded(spec, out) {
+    const t = classifyPath(spec);
+    if (t === "file") {
+        out.add(spec);
+        return;
+    }
+    // Glob or directory — expand.
+    const files = await expandGlobs([spec]);
+    for (const f of files)
+        out.add(f);
+}
+export function resolveFileSource(p) {
+    const abs = resolvePath(p);
+    return {
+        source: { id: abs, type: "file" },
+        content: null, // let rg read the file directly (streaming, no temp file)
+    };
+}
+export async function resolveGlobSource(pattern) {
+    const files = await expandGlobs([pattern]);
+    return files.map((f) => ({
+        source: { id: resolvePath(f), type: "file" },
+        content: null,
+    }));
+}
+export function resolveCommandSource(cmd) {
+    return {
+        source: { id: `cmd:${cmd}`, type: "command", label: `$ ${cmd}` },
+        content: null, // will be filled by captureCommand
+    };
+}
+/** Cap captured command stdout at 64MB. Past that we truncate with a marker. */
+const COMMAND_OUTPUT_MAX_BYTES = 64 * 1024 * 1024;
+/** Default command timeout — 60s for `git log` etc. is plenty. */
+const COMMAND_TIMEOUT_MS = 60_000;
+/**
+ * Capture a shell command's stdout for searching.
+ *
+ * Quoting handled correctly: the command runs through the platform
+ * shell (`bash -c` on POSIX, `cmd /c` on Windows), so `git log
+ * --grep="fix bug"` parses the way the user typed it.
+ *
+ * Output is capped at COMMAND_OUTPUT_MAX_BYTES and the command is
+ * killed after COMMAND_TIMEOUT_MS so a hanging or runaway command
+ * can't lock up the agent harness.
+ */
+export async function captureCommand(cmd) {
+    const trimmed = cmd.trim();
+    if (!trimmed)
+        throw new Error("Empty command");
+    const shell = process.platform === "win32" ? "cmd.exe" : "bash";
+    const shellArgs = process.platform === "win32" ? ["/c", trimmed] : ["-c", trimmed];
+    return await new Promise((resolve, reject) => {
+        const proc = spawn(shell, shellArgs, { stdio: ["ignore", "pipe", "pipe"] });
+        const chunks = [];
+        let bytes = 0;
+        let truncated = false;
+        let stderr = "";
+        let settled = false;
+        const timer = setTimeout(() => {
+            if (settled)
+                return;
+            settled = true;
+            try {
+                proc.kill("SIGTERM");
+            }
+            catch { /* ignore */ }
+            reject(new Error(`Command timed out after ${COMMAND_TIMEOUT_MS}ms: ${trimmed.slice(0, 200)}`));
+        }, COMMAND_TIMEOUT_MS);
+        proc.stdout.on("data", (chunk) => {
+            if (truncated)
+                return;
+            const remaining = COMMAND_OUTPUT_MAX_BYTES - bytes;
+            if (chunk.length <= remaining) {
+                chunks.push(chunk);
+                bytes += chunk.length;
+                return;
+            }
+            // Take what we can, then signal SIGTERM. Anything after the cap
+            // is silently dropped — we keep a marker so the caller can tell.
+            if (remaining > 0) {
+                chunks.push(chunk.subarray(0, remaining));
+                bytes += remaining;
+            }
+            truncated = true;
+            try {
+                proc.kill("SIGTERM");
+            }
+            catch { /* ignore */ }
+        });
+        proc.stderr.setEncoding("utf8");
+        proc.stderr.on("data", (chunk) => { stderr += chunk; });
+        proc.on("error", (err) => {
+            if (settled)
+                return;
+            settled = true;
+            clearTimeout(timer);
+            reject(err);
+        });
+        proc.on("close", (code) => {
+            if (settled)
+                return;
+            settled = true;
+            clearTimeout(timer);
+            if (truncated) {
+                // Successful path: caller gets the truncated output plus a
+                // marker. We don't reject because partial data is still useful.
+                const out = Buffer.concat(chunks).toString("utf8") +
+                    `\n[mpg: command output truncated at ${COMMAND_OUTPUT_MAX_BYTES} bytes]\n`;
+                resolve(out);
+                return;
+            }
+            if (code !== 0 && code !== null) {
+                reject(new Error(`Command exited with code ${code}: ${trimmed.slice(0, 200)}` +
+                    (stderr ? `\nstderr: ${stderr.slice(0, 500)}` : "")));
+                return;
+            }
+            resolve(Buffer.concat(chunks).toString("utf8"));
+        });
+    });
+}
+/** Deprecated: use getStdin() instead to avoid double-reads. */
+export async function captureStdin() {
+    return getStdin();
+}
+/** Cap fetched URL body at 16MB — anything larger is a denial-of-context risk. */
+const URL_FETCH_MAX_BYTES = 16 * 1024 * 1024;
+/** Default URL fetch timeout. */
+const URL_FETCH_TIMEOUT_MS = 30_000;
+export async function captureUrl(url) {
+    const controller = new AbortController();
+    const timer = setTimeout(() => controller.abort(), URL_FETCH_TIMEOUT_MS);
+    let res;
+    try {
+        res = await fetch(url, {
+            redirect: "follow",
+            headers: { "user-agent": "mpg/0.3 (+https://github.com/JadeZaher/mpg)" },
+            signal: controller.signal,
+        });
+    }
+    catch (err) {
+        clearTimeout(timer);
+        if (err.name === "AbortError") {
+            throw new Error(`Fetch of ${url} timed out after ${URL_FETCH_TIMEOUT_MS}ms`);
+        }
+        throw err;
+    }
+    try {
+        if (!res.ok) {
+            throw new Error(`Failed to fetch ${url}: ${res.status} ${res.statusText}`);
+        }
+        // Cheap MIME guard — we are searching text. Reject obvious binary
+        // types before we read the body so an LLM can't OOM us by passing
+        // a video URL.
+        const ct = (res.headers.get("content-type") ?? "").toLowerCase();
+        if (ct && !ct.startsWith("text/") && !/json|xml|yaml|javascript|csv|html|markdown/.test(ct)) {
+            throw new Error(`Refusing to fetch non-text content-type "${ct}" from ${url}. ` +
+                `Use a different tool to search binary payloads.`);
+        }
+        // Content-length pre-check (cheap if the server set it).
+        const clHeader = res.headers.get("content-length");
+        if (clHeader) {
+            const cl = parseInt(clHeader, 10);
+            if (!Number.isNaN(cl) && cl > URL_FETCH_MAX_BYTES) {
+                throw new Error(`Refusing to fetch ${cl} bytes from ${url} (cap: ${URL_FETCH_MAX_BYTES}). ` +
+                    `Download manually and search the file.`);
+            }
+        }
+        // Stream-with-cap. If content-length lied or wasn't set, we still
+        // bail out as soon as we cross the threshold.
+        if (!res.body)
+            return "";
+        const reader = res.body.getReader();
+        const decoder = new TextDecoder("utf-8");
+        let out = "";
+        let bytes = 0;
+        while (true) {
+            const { value, done } = await reader.read();
+            if (done)
+                break;
+            bytes += value.byteLength;
+            if (bytes > URL_FETCH_MAX_BYTES) {
+                try {
+                    await reader.cancel();
+                }
+                catch { /* ignore */ }
+                throw new Error(`Fetched body exceeded ${URL_FETCH_MAX_BYTES} bytes from ${url}. ` +
+                    `Download manually and search the file.`);
+            }
+            out += decoder.decode(value, { stream: true });
+        }
+        out += decoder.decode();
+        return out;
+    }
+    finally {
+        clearTimeout(timer);
+    }
+}
+export function resolveUrlSource(url) {
+    return {
+        source: { id: url, type: "url" },
+        content: null, // filled by captureUrl
+    };
+}
+//# sourceMappingURL=sources.js.map

package/dist/sources.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"sources.js","sourceRoot":"","sources":["../src/sources.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AAEH,OAAO,EAAE,KAAK,EAAE,MAAM,oBAAoB,CAAC;AAC3C,OAAO,EAAE,UAAU,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAC;AAC/C,OAAO,EAAE,OAAO,IAAI,WAAW,EAAE,MAAM,WAAW,CAAC;AAInD,sEAAsE;AACtE,IAAI,YAAY,GAAkB,IAAI,CAAC;AAEvC,8EAA8E;AAC9E,MAAM,CAAC,KAAK,UAAU,QAAQ;IAC5B,IAAI,YAAY,KAAK,IAAI;QAAE,OAAO,YAAY,CAAC;IAC/C,IAAI,OAAO,CAAC,KAAK,CAAC,KAAK,EAAE,CAAC;QACxB,YAAY,GAAG,EAAE,CAAC;QAClB,OAAO,EAAE,CAAC;IACZ,CAAC;IACD,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,OAAO,CAAC,KAAiB,EAAE,CAAC;QACpD,MAAM,CAAC,IAAI,CAAC,OAAO,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;IACtE,CAAC;IACD,YAAY,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;IACtD,OAAO,YAAY,CAAC;AACtB,CAAC;AAED,sDAAsD;AACtD,MAAM,UAAU,eAAe;IAC7B,YAAY,GAAG,IAAI,CAAC;AACtB,CAAC;AAQD,6EAA6E;AAC7E,MAAM,CAAC,KAAK,UAAU,WAAW,CAAC,QAAkB;IAClD,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IACrC,MAAM,EAAE,IAAI,EAAE,GAAG,MAAM,MAAM,CAAC,kBAAkB,CAAC,CAAC;IAClD,MAAM,EAAE,QAAQ,EAAE,GAAG,MAAM,MAAM,CAAC,SAAS,CAAC,CAAC;IAC7C,MAAM,GAAG,GAAG,IAAI,GAAG,EAAU,CAAC;IAE9B,KAAK,UAAU,IAAI,CAAC,CAAS;QAC3B,4DAA4D;QAC5D,0BAA0B;QAC1B,oCAAoC;QACpC,IAAI,CAAC,GAAmC,IAAI,CAAC;QAC7C,IAAI,CAAC;YAAC,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;QAAC,CAAC;QAAC,MAAM,CAAC,CAAC,iBAAiB,CAAC,CAAC;QAEpD,kEAAkE;QAClE,mEAAmE;QACnE,iEAAiE;QACjE,MAAM,aAAa,GAAG,CAAC,CAAS,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC;QAE3D,IAAI,CAAC,IAAI,CAAC,CAAC,WAAW,EAAE,EAAE,CAAC;YACzB,MAAM,OAAO,GAAG,CAAC,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC;YAC1C,MAAM,cAAc,GAAG,GAAG,aAAa,CAAC,OAAO,CAAC,KAAK,CAAC;YACtD,IAAI,CAAC;gBACH,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,IAAI,CAAC,cAAc,CAAC,EAAE,CAAC;oBAC/C,IAAI,CAAC;wBACH,MAAM,EAAE,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC;wBAC3B,IAAI,EAAE,CAAC,MAAM,EAAE;4BAAE,GAAG,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;oBAClC,CAAC;oBAAC,MAAM,CAAC,CAAC,UAAU,CAAC,CAAC;gBACxB,CAAC;YACH,CAAC;YAAC,MAAM,CAAC,CAAC,YAAY,CAAC,CAAC;QAC1B,CAAC;aAAM,IAAI,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC;YAC3B,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;QACb,CAAC;aAAM,CAAC;YACN,yCAAyC;YACzC,IAAI,CAAC;gBACH,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,IAAI,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;oBACjD,IAAI,CAAC;wBACH,MAAM,EAAE,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC;wBAC3B,IAAI,EAAE,CAAC,MAAM,EAAE;4BAAE,GAAG,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;oBAClC,CAAC;oBAAC,MAAM,CAAC,CAAC,UAAU,CAAC,CAAC;gBACxB,CAAC;YACH,CAAC;YAAC,MAAM,CAAC,CAAC,YAAY,CAAC,CAAC;QAC1B,CAAC;IACH,CAAC;IAED,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;QAC/B,MAAM,IAAI,CAAC,OAAO,CAAC,CAAC;IACtB,CAAC;IACD,OAAO,CAAC,GAAG,GAAG,CAAC,CAAC;AAClB,CAAC;AAED,8DAA8D;AAC9D,MAAM,UAAU,YAAY,CAAC,CAAS;IACpC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC;QAAE,OAAO,MAAM,CAAC;IAClC,MAAM,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;IACtB,IAAI,CAAC,CAAC,WAAW,EAAE;QAAE,OAAO,MAAM,CAAC;IACnC,IAAI,CAAC,CAAC,MAAM,EAAE;QAAE,OAAO,MAAM,CAAC;IAC9B,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;;;;;;;;;;;;;;;GAgBG;AACH,MAAM,CAAC,KAAK,UAAU,gBAAgB,CAAC,KAAe,EAAE,YAA4B;IAClF,MAAM,GAAG,GAAG,IAAI,GAAG,EAAU,CAAC;IAC9B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;YAClB,yBAAyB;YACzB,MAAM,IAAI,GAAG,YAAY,IAAI,MAAM,QAAQ,EAAE,CAAC;YAC9C,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC;gBACvC,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;gBAC5B,IAAI,CAAC,OAAO,IAAI,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC;oBAAE,SAAS;gBAClD,MAAM,WAAW,CAAC,OAAO,EAAE,GAAG,CAAC,CAAC;YAClC,CAAC;YACD,SAAS;QACX,CAAC;QACD,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;YACzB,0BAA0B;YAC1B,MAAM,QAAQ,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;YAC/B,IAAI,IAAY,CAAC;YACjB,IAAI,CAAC;gBACH,MAAM,EAAE,YAAY,EAAE,GAAG,MAAM,MAAM,CAAC,SAAS,CAAC,CAAC;gBACjD,IAAI,GAAG,YAAY,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAC;YACxC,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,MAAM,IAAI,KAAK,CAAC,+BAA+B,QAAQ,KAAM,GAAa,CAAC,OAAO,EAAE,CAAC,CAAC;YACxF,CAAC;YACD,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC;gBACvC,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;gBAC5B,IAAI,CAAC,OAAO,IAAI,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC;oBAAE,SAAS;gBAClD,MAAM,WAAW,CAAC,OAAO,EAAE,GAAG,CAAC,CAAC;YAClC,CAAC;YACD,SAAS;QACX,CAAC;QACD,MAAM,WAAW,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;IAC/B,CAAC;IACD,OAAO,CAAC,GAAG,GAAG,CAAC,CAAC;AAClB,CAAC;AAED;;;;;;;;;;;;;;;;;GAiBG;AACH,MAAM,CAAC,KAAK,UAAU,iBAAiB,CACrC,KAAe,EACf,YAA4B;IAE5B,MAAM,KAAK,GAAG,IAAI,GAAG,EAAU,CAAC;IAChC,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;IAE/B,SAAS,WAAW,CAAC,CAAS;QAC5B,+DAA+D;QAC/D,mEAAmE;QACnE,gCAAgC;QAChC,OAAO,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAC5B,CAAC;IAED,KAAK,UAAU,QAAQ,CAAC,IAAY;QAClC,IAAI,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC;YACrB,MAAM,CAAC,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC;YACzB,IAAI,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC;gBACf,KAAK,CAAC,GAAG,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC,CAAC;gBAC7B,OAAO;YACT,CAAC;YACD,IAAI,CAAC,CAAC,WAAW,EAAE,EAAE,CAAC;gBACpB,yDAAyD;gBACzD,6DAA6D;gBAC7D,kBAAkB;gBAClB,IAAI,CAAC,GAAG,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC,CAAC;gBAC5B,OAAO;YACT,CAAC;QACH,CAAC;QACD,IAAI,WAAW,CAAC,IAAI,CAAC,EAAE,CAAC;YACtB,4DAA4D;YAC5D,6DAA6D;YAC7D,0DAA0D;YAC1D,mCAAmC;YACnC,MAAM,QAAQ,GAAG,MAAM,WAAW,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;YAC3C,KAAK,MAAM,CAAC,IAAI,QAAQ;gBAAE,KAAK,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,CAAC;YACpD,OAAO;QACT,CAAC;QACD,kEAAkE;QAClE,4DAA4D;QAC5D,0BAA0B;QAC1B,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;IACjB,CAAC;IAED,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;YAClB,MAAM,IAAI,GAAG,YAAY,IAAI,MAAM,QAAQ,EAAE,CAAC;YAC9C,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC;gBACvC,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;gBAC5B,IAAI,CAAC,OAAO,IAAI,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC;oBAAE,SAAS;gBAClD,MAAM,QAAQ,CAAC,OAAO,CAAC,CAAC;YAC1B,CAAC;YACD,SAAS;QACX,CAAC;QACD,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;YACzB,MAAM,QAAQ,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;YAC/B,IAAI,IAAY,CAAC;YACjB,IAAI,CAAC;gBACH,MAAM,EAAE,YAAY,EAAE,GAAG,MAAM,MAAM,CAAC,SAAS,CAAC,CAAC;gBACjD,IAAI,GAAG,YAAY,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAC;YACxC,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,MAAM,IAAI,KAAK,CAAC,+BAA+B,QAAQ,KAAM,GAAa,CAAC,OAAO,EAAE,CAAC,CAAC;YACxF,CAAC;YACD,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC;gBACvC,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;gBAC5B,IAAI,CAAC,OAAO,IAAI,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC;oBAAE,SAAS;gBAClD,MAAM,QAAQ,CAAC,OAAO,CAAC,CAAC;YAC1B,CAAC;YACD,SAAS;QACX,CAAC;QACD,MAAM,QAAQ,CAAC,IAAI,CAAC,CAAC;IACvB,CAAC;IACD,OAAO,EAAE,KAAK,EAAE,CAAC,GAAG,KAAK,CAAC,EAAE,IAAI,EAAE,CAAC,GAAG,IAAI,CAAC,EAAE,CAAC;AAChD,CAAC;AAED,iEAAiE;AACjE,KAAK,UAAU,WAAW,CAAC,IAAY,EAAE,GAAgB;IACvD,MAAM,CAAC,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC;IAC7B,IAAI,CAAC,KAAK,MAAM,EAAE,CAAC;QACjB,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QACd,OAAO;IACT,CAAC;IACD,8BAA8B;IAC9B,MAAM,KAAK,GAAG,MAAM,WAAW,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IACxC,KAAK,MAAM,CAAC,IAAI,KAAK;QAAE,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;AACpC,CAAC;AAED,MAAM,UAAU,iBAAiB,CAAC,CAAS;IACzC,MAAM,GAAG,GAAG,WAAW,CAAC,CAAC,CAAC,CAAC;IAC3B,OAAO;QACL,MAAM,EAAE,EAAE,EAAE,EAAE,GAAG,EAAE,IAAI,EAAE,MAAM,EAAE;QACjC,OAAO,EAAE,IAAI,EAAE,0DAA0D;KAC1E,CAAC;AACJ,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,iBAAiB,CAAC,OAAe;IACrD,MAAM,KAAK,GAAG,MAAM,WAAW,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC;IAC3C,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;QACvB,MAAM,EAAE,EAAE,EAAE,EAAE,WAAW,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE;QAC5C,OAAO,EAAE,IAAI;KACd,CAAC,CAAC,CAAC;AACN,CAAC;AAED,MAAM,UAAU,oBAAoB,CAAC,GAAW;IAC9C,OAAO;QACL,MAAM,EAAE,EAAE,EAAE,EAAE,OAAO,GAAG,EAAE,EAAE,IAAI,EAAE,SAAS,EAAE,KAAK,EAAE,KAAK,GAAG,EAAE,EAAE;QAChE,OAAO,EAAE,IAAI,EAAE,mCAAmC;KACnD,CAAC;AACJ,CAAC;AAED,gFAAgF;AAChF,MAAM,wBAAwB,GAAG,EAAE,GAAG,IAAI,GAAG,IAAI,CAAC;AAClD,kEAAkE;AAClE,MAAM,kBAAkB,GAAG,MAAM,CAAC;AAElC;;;;;;;;;;GAUG;AACH,MAAM,CAAC,KAAK,UAAU,cAAc,CAAC,GAAW;IAC9C,MAAM,OAAO,GAAG,GAAG,CAAC,IAAI,EAAE,CAAC;IAC3B,IAAI,CAAC,OAAO;QAAE,MAAM,IAAI,KAAK,CAAC,eAAe,CAAC,CAAC;IAE/C,MAAM,KAAK,GAAG,OAAO,CAAC,QAAQ,KAAK,OAAO,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC;IAChE,MAAM,SAAS,GAAG,OAAO,CAAC,QAAQ,KAAK,OAAO,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;IAEnF,OAAO,MAAM,IAAI,OAAO,CAAS,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;QACnD,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,EAAE,SAAS,EAAE,EAAE,KAAK,EAAE,CAAC,QAAQ,EAAE,MAAM,EAAE,MAAM,CAAC,EAAE,CAAC,CAAC;QAC5E,MAAM,MAAM,GAAa,EAAE,CAAC;QAC5B,IAAI,KAAK,GAAG,CAAC,CAAC;QACd,IAAI,SAAS,GAAG,KAAK,CAAC;QACtB,IAAI,MAAM,GAAG,EAAE,CAAC;QAChB,IAAI,OAAO,GAAG,KAAK,CAAC;QAEpB,MAAM,KAAK,GAAG,UAAU,CAAC,GAAG,EAAE;YAC5B,IAAI,OAAO;gBAAE,OAAO;YACpB,OAAO,GAAG,IAAI,CAAC;YACf,IAAI,CAAC;gBAAC,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;YAAC,CAAC;YAAC,MAAM,CAAC,CAAC,YAAY,CAAC,CAAC;YACpD,MAAM,CAAC,IAAI,KAAK,CACd,2BAA2B,kBAAkB,OAAO,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAC5E,CAAC,CAAC;QACL,CAAC,EAAE,kBAAkB,CAAC,CAAC;QAEvB,IAAI,CAAC,MAAO,CAAC,EAAE,CAAC,MAAM,EAAE,CAAC,KAAa,EAAE,EAAE;YACxC,IAAI,SAAS;gBAAE,OAAO;YACtB,MAAM,SAAS,GAAG,wBAAwB,GAAG,KAAK,CAAC;YACnD,IAAI,KAAK,CAAC,MAAM,IAAI,SAAS,EAAE,CAAC;gBAC9B,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;gBACnB,KAAK,IAAI,KAAK,CAAC,MAAM,CAAC;gBACtB,OAAO;YACT,CAAC;YACD,gEAAgE;YAChE,iEAAiE;YACjE,IAAI,SAAS,GAAG,CAAC,EAAE,CAAC;gBAClB,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC,CAAC;gBAC1C,KAAK,IAAI,SAAS,CAAC;YACrB,CAAC;YACD,SAAS,GAAG,IAAI,CAAC;YACjB,IAAI,CAAC;gBAAC,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;YAAC,CAAC;YAAC,MAAM,CAAC,CAAC,YAAY,CAAC,CAAC;QACtD,CAAC,CAAC,CAAC;QAEH,IAAI,CAAC,MAAO,CAAC,WAAW,CAAC,MAAM,CAAC,CAAC;QACjC,IAAI,CAAC,MAAO,CAAC,EAAE,CAAC,MAAM,EAAE,CAAC,KAAa,EAAE,EAAE,GAAG,MAAM,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QAEjE,IAAI,CAAC,EAAE,CAAC,OAAO,EAAE,CAAC,GAAG,EAAE,EAAE;YACvB,IAAI,OAAO;gBAAE,OAAO;YACpB,OAAO,GAAG,IAAI,CAAC;YACf,YAAY,CAAC,KAAK,CAAC,CAAC;YACpB,MAAM,CAAC,GAAG,CAAC,CAAC;QACd,CAAC,CAAC,CAAC;QAEH,IAAI,CAAC,EAAE,CAAC,OAAO,EAAE,CAAC,IAAI,EAAE,EAAE;YACxB,IAAI,OAAO;gBAAE,OAAO;YACpB,OAAO,GAAG,IAAI,CAAC;YACf,YAAY,CAAC,KAAK,CAAC,CAAC;YACpB,IAAI,SAAS,EAAE,CAAC;gBACd,2DAA2D;gBAC3D,gEAAgE;gBAChE,MAAM,GAAG,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,QAAQ,CAAC,MAAM,CAAC;oBAChD,uCAAuC,wBAAwB,WAAW,CAAC;gBAC7E,OAAO,CAAC,GAAG,CAAC,CAAC;gBACb,OAAO;YACT,CAAC;YACD,IAAI,IAAI,KAAK,CAAC,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;gBAChC,MAAM,CAAC,IAAI,KAAK,CACd,4BAA4B,IAAI,KAAK,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE;oBAC5D,CAAC,MAAM,CAAC,CAAC,CAAC,aAAa,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CACpD,CAAC,CAAC;gBACH,OAAO;YACT,CAAC;YACD,OAAO,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC;QAClD,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;AACL,CAAC;AAED,gEAAgE;AAChE,MAAM,CAAC,KAAK,UAAU,YAAY;IAChC,OAAO,QAAQ,EAAE,CAAC;AACpB,CAAC;AAED,kFAAkF;AAClF,MAAM,mBAAmB,GAAG,EAAE,GAAG,IAAI,GAAG,IAAI,CAAC;AAC7C,iCAAiC;AACjC,MAAM,oBAAoB,GAAG,MAAM,CAAC;AAEpC,MAAM,CAAC,KAAK,UAAU,UAAU,CAAC,GAAW;IAC1C,MAAM,UAAU,GAAG,IAAI,eAAe,EAAE,CAAC;IACzC,MAAM,KAAK,GAAG,UAAU,CAAC,GAAG,EAAE,CAAC,UAAU,CAAC,KAAK,EAAE,EAAE,oBAAoB,CAAC,CAAC;IACzE,IAAI,GAAa,CAAC;IAClB,IAAI,CAAC;QACH,GAAG,GAAG,MAAM,KAAK,CAAC,GAAG,EAAE;YACrB,QAAQ,EAAE,QAAQ;YAClB,OAAO,EAAE,EAAE,YAAY,EAAE,6CAA6C,EAAE;YACxE,MAAM,EAAE,UAAU,CAAC,MAAM;SAC1B,CAAC,CAAC;IACL,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,YAAY,CAAC,KAAK,CAAC,CAAC;QACpB,IAAK,GAAa,CAAC,IAAI,KAAK,YAAY,EAAE,CAAC;YACzC,MAAM,IAAI,KAAK,CAAC,YAAY,GAAG,oBAAoB,oBAAoB,IAAI,CAAC,CAAC;QAC/E,CAAC;QACD,MAAM,GAAG,CAAC;IACZ,CAAC;IACD,IAAI,CAAC;QACH,IAAI,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC;YACZ,MAAM,IAAI,KAAK,CAAC,mBAAmB,GAAG,KAAK,GAAG,CAAC,MAAM,IAAI,GAAG,CAAC,UAAU,EAAE,CAAC,CAAC;QAC7E,CAAC;QACD,kEAAkE;QAClE,kEAAkE;QAClE,eAAe;QACf,MAAM,EAAE,GAAG,CAAC,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;QACjE,IAAI,EAAE,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,OAAO,CAAC,IAAI,CAAC,4CAA4C,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,CAAC;YAC5F,MAAM,IAAI,KAAK,CACb,4CAA4C,EAAE,UAAU,GAAG,IAAI;gBAC/D,iDAAiD,CAClD,CAAC;QACJ,CAAC;QACD,yDAAyD;QACzD,MAAM,QAAQ,GAAG,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,gBAAgB,CAAC,CAAC;QACnD,IAAI,QAAQ,EAAE,CAAC;YACb,MAAM,EAAE,GAAG,QAAQ,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;YAClC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,EAAE,GAAG,mBAAmB,EAAE,CAAC;gBAClD,MAAM,IAAI,KAAK,CACb,qBAAqB,EAAE,eAAe,GAAG,UAAU,mBAAmB,KAAK;oBAC3E,wCAAwC,CACzC,CAAC;YACJ,CAAC;QACH,CAAC;QACD,kEAAkE;QAClE,8CAA8C;QAC9C,IAAI,CAAC,GAAG,CAAC,IAAI;YAAE,OAAO,EAAE,CAAC;QACzB,MAAM,MAAM,GAAG,GAAG,CAAC,IAAI,CAAC,SAAS,EAAE,CAAC;QACpC,MAAM,OAAO,GAAG,IAAI,WAAW,CAAC,OAAO,CAAC,CAAC;QACzC,IAAI,GAAG,GAAG,EAAE,CAAC;QACb,IAAI,KAAK,GAAG,CAAC,CAAC;QACd,OAAO,IAAI,EAAE,CAAC;YACZ,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,GAAG,MAAM,MAAM,CAAC,IAAI,EAAE,CAAC;YAC5C,IAAI,IAAI;gBAAE,MAAM;YAChB,KAAK,IAAI,KAAK,CAAC,UAAU,CAAC;YAC1B,IAAI,KAAK,GAAG,mBAAmB,EAAE,CAAC;gBAChC,IAAI,CAAC;oBAAC,MAAM,MAAM,CAAC,MAAM,EAAE,CAAC;gBAAC,CAAC;gBAAC,MAAM,CAAC,CAAC,YAAY,CAAC,CAAC;gBACrD,MAAM,IAAI,KAAK,CACb,yBAAyB,mBAAmB,eAAe,GAAG,IAAI;oBAClE,wCAAwC,CACzC,CAAC;YACJ,CAAC;YACD,GAAG,IAAI,OAAO,CAAC,MAAM,CAAC,KAAK,EAAE,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC;QACjD,CAAC;QACD,GAAG,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;QACxB,OAAO,GAAG,CAAC;IACb,CAAC;YAAS,CAAC;QACT,YAAY,CAAC,KAAK,CAAC,CAAC;IACtB,CAAC;AACH,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,GAAW;IAC1C,OAAO;QACL,MAAM,EAAE,EAAE,EAAE,EAAE,GAAG,EAAE,IAAI,EAAE,KAAK,EAAE;QAChC,OAAO,EAAE,IAAI,EAAE,uBAAuB;KACvC,CAAC;AACJ,CAAC"}

package/dist/tokens.d.ts

@@ -0,0 +1,35 @@
+/**
+ * Token estimation.
+ *
+ * We use a simple chars/4 heuristic as the default. It's an approximation:
+ * - English prose averages ~4 chars/token
+ * - Code averages ~3.5 chars/token (more symbols)
+ * - JSON averages ~3 chars/token
+ *
+ * This is intentionally dependency-free. For higher fidelity, callers
+ * can plug in tiktoken or gpt-tokenizer at the integration boundary.
+ *
+ * The heuristic is good enough to make *budgeting* decisions (sizing
+ * context windows, capping output) which is what mpg is for. It's not
+ * a substitute for a real tokenizer when billing accuracy matters.
+ */
+export interface TokenModel {
+    /** Estimate tokens for a single string. */
+    estimate(text: string): number;
+    /** Estimate tokens across an array of strings (slightly faster, no array alloc). */
+    estimateMany(texts: string[]): number;
+}
+export declare const defaultTokens: TokenModel;
+/**
+ * Trim a list of lines to fit within a token budget, preferring the
+ * lines closest to a target index.
+ *
+ * This is the workhorse used to build the pre/post context windows of
+ * a node. It walks outward from the match line, accumulating lines
+ * until the budget is exhausted, then returns the kept lines in
+ * original order.
+ */
+export declare function trimLinesToBudget(lines: string[], targetIndex: number, budget: number, model?: TokenModel): {
+    kept: string[];
+    spent: number;
+};

package/dist/tokens.js

@@ -0,0 +1,95 @@
+/**
+ * Token estimation.
+ *
+ * We use a simple chars/4 heuristic as the default. It's an approximation:
+ * - English prose averages ~4 chars/token
+ * - Code averages ~3.5 chars/token (more symbols)
+ * - JSON averages ~3 chars/token
+ *
+ * This is intentionally dependency-free. For higher fidelity, callers
+ * can plug in tiktoken or gpt-tokenizer at the integration boundary.
+ *
+ * The heuristic is good enough to make *budgeting* decisions (sizing
+ * context windows, capping output) which is what mpg is for. It's not
+ * a substitute for a real tokenizer when billing accuracy matters.
+ */
+const DEFAULT_CHARS_PER_TOKEN = 4;
+class HeuristicTokenModel {
+    charsPerToken;
+    constructor(charsPerToken = DEFAULT_CHARS_PER_TOKEN) {
+        this.charsPerToken = charsPerToken;
+    }
+    estimate(text) {
+        if (!text)
+            return 0;
+        // Round up so empty/single-char strings get 1 token, not 0.
+        return Math.max(1, Math.ceil(text.length / this.charsPerToken));
+    }
+    estimateMany(texts) {
+        let total = 0;
+        for (const t of texts)
+            total += this.estimate(t);
+        return total;
+    }
+}
+export const defaultTokens = new HeuristicTokenModel();
+/**
+ * Trim a list of lines to fit within a token budget, preferring the
+ * lines closest to a target index.
+ *
+ * This is the workhorse used to build the pre/post context windows of
+ * a node. It walks outward from the match line, accumulating lines
+ * until the budget is exhausted, then returns the kept lines in
+ * original order.
+ */
+export function trimLinesToBudget(lines, targetIndex, budget, model = defaultTokens) {
+    if (budget <= 0 || lines.length === 0) {
+        return { kept: [], spent: 0 };
+    }
+    // Greedy outward expansion: keep a window of lines around targetIndex
+    // that fits in `budget` tokens.
+    const kept = new Array(lines.length).fill(null);
+    kept[targetIndex] = lines[targetIndex];
+    let spent = model.estimate(lines[targetIndex]);
+    let lo = targetIndex - 1;
+    let hi = targetIndex + 1;
+    // Alternate between expanding above and below, preferring whichever
+    // side has more remaining lines. This produces a balanced window.
+    while (lo >= 0 || hi < lines.length) {
+        const canLo = lo >= 0;
+        const canHi = hi < lines.length;
+        if (!canLo && !canHi)
+            break;
+        // Prefer the side with more remaining lines (balanced growth).
+        const takeLo = canLo && (!canHi || (targetIndex - lo) <= (hi - targetIndex));
+        if (takeLo) {
+            const cost = model.estimate(lines[lo]);
+            if (spent + cost > budget) {
+                lo = -1; // stop expanding up
+            }
+            else {
+                kept[lo] = lines[lo];
+                spent += cost;
+                lo--;
+            }
+        }
+        else {
+            const cost = model.estimate(lines[hi]);
+            if (spent + cost > budget) {
+                hi = lines.length; // stop expanding down
+            }
+            else {
+                kept[hi] = lines[hi];
+                spent += cost;
+                hi++;
+            }
+        }
+    }
+    const result = [];
+    for (let i = 0; i < kept.length; i++) {
+        if (kept[i] !== null)
+            result.push(kept[i]);
+    }
+    return { kept: result, spent };
+}
+//# sourceMappingURL=tokens.js.map

package/dist/tokens.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"tokens.js","sourceRoot":"","sources":["../src/tokens.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AAEH,MAAM,uBAAuB,GAAG,CAAC,CAAC;AASlC,MAAM,mBAAmB;IACH;IAApB,YAAoB,gBAAwB,uBAAuB;QAA/C,kBAAa,GAAb,aAAa,CAAkC;IAAG,CAAC;IAEvE,QAAQ,CAAC,IAAY;QACnB,IAAI,CAAC,IAAI;YAAE,OAAO,CAAC,CAAC;QACpB,4DAA4D;QAC5D,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC,aAAa,CAAC,CAAC,CAAC;IAClE,CAAC;IAED,YAAY,CAAC,KAAe;QAC1B,IAAI,KAAK,GAAG,CAAC,CAAC;QACd,KAAK,MAAM,CAAC,IAAI,KAAK;YAAE,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC;QACjD,OAAO,KAAK,CAAC;IACf,CAAC;CACF;AAED,MAAM,CAAC,MAAM,aAAa,GAAe,IAAI,mBAAmB,EAAE,CAAC;AAEnE;;;;;;;;GAQG;AACH,MAAM,UAAU,iBAAiB,CAC/B,KAAe,EACf,WAAmB,EACnB,MAAc,EACd,QAAoB,aAAa;IAEjC,IAAI,MAAM,IAAI,CAAC,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACtC,OAAO,EAAE,IAAI,EAAE,EAAE,EAAE,KAAK,EAAE,CAAC,EAAE,CAAC;IAChC,CAAC;IAED,sEAAsE;IACtE,gCAAgC;IAChC,MAAM,IAAI,GAAG,IAAI,KAAK,CAAgB,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC/D,IAAI,CAAC,WAAW,CAAC,GAAG,KAAK,CAAC,WAAW,CAAC,CAAC;IACvC,IAAI,KAAK,GAAG,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC,WAAW,CAAC,CAAC,CAAC;IAE/C,IAAI,EAAE,GAAG,WAAW,GAAG,CAAC,CAAC;IACzB,IAAI,EAAE,GAAG,WAAW,GAAG,CAAC,CAAC;IAEzB,oEAAoE;IACpE,kEAAkE;IAClE,OAAO,EAAE,IAAI,CAAC,IAAI,EAAE,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;QACpC,MAAM,KAAK,GAAG,EAAE,IAAI,CAAC,CAAC;QACtB,MAAM,KAAK,GAAG,EAAE,GAAG,KAAK,CAAC,MAAM,CAAC;QAEhC,IAAI,CAAC,KAAK,IAAI,CAAC,KAAK;YAAE,MAAM;QAE5B,+DAA+D;QAC/D,MAAM,MAAM,GACV,KAAK,IAAI,CAAC,CAAC,KAAK,IAAI,CAAC,WAAW,GAAG,EAAE,CAAC,IAAI,CAAC,EAAE,GAAG,WAAW,CAAC,CAAC,CAAC;QAEhE,IAAI,MAAM,EAAE,CAAC;YACX,MAAM,IAAI,GAAG,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC;YACvC,IAAI,KAAK,GAAG,IAAI,GAAG,MAAM,EAAE,CAAC;gBAC1B,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,oBAAoB;YAC/B,CAAC;iBAAM,CAAC;gBACN,IAAI,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,EAAE,CAAC,CAAC;gBACrB,KAAK,IAAI,IAAI,CAAC;gBACd,EAAE,EAAE,CAAC;YACP,CAAC;QACH,CAAC;aAAM,CAAC;YACN,MAAM,IAAI,GAAG,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC;YACvC,IAAI,KAAK,GAAG,IAAI,GAAG,MAAM,EAAE,CAAC;gBAC1B,EAAE,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,sBAAsB;YAC3C,CAAC;iBAAM,CAAC;gBACN,IAAI,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,EAAE,CAAC,CAAC;gBACrB,KAAK,IAAI,IAAI,CAAC;gBACd,EAAE,EAAE,CAAC;YACP,CAAC;QACH,CAAC;IACH,CAAC;IAED,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACrC,IAAI,IAAI,CAAC,CAAC,CAAC,KAAK,IAAI;YAAE,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAW,CAAC,CAAC;IACvD,CAAC;IACD,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC;AACjC,CAAC"}