@mrclrchtr/supi-code-intelligence 1.7.0 → 1.8.0

package/src/actions/pattern-action.ts

@@ -1,392 +0,0 @@
-// Pattern action — bounded, scope-aware text search.
-// biome-ignore-all lint/nursery/noExcessiveLinesPerFile: text and structured pattern flows share one formatter and matcher pipeline
-
-import {
-  getStructuredPatternMatches,
-  isStructuredPatternKind,
-  type StructuredMatch,
-  type StructuredPatternResult,
-} from "../pattern-structured.ts";
-import type { CodeQueryParams as ActionParams } from "../query-params.ts";
-import type { RgMatch } from "../search-helpers.ts";
-import {
-  escapeRegex,
-  groupByFile,
-  normalizePath,
-  runRipgrep,
-  runRipgrepDetailed,
-} from "../search-helpers.ts";
-import type { CodeIntelResult, SearchDetails } from "../types.ts";
-
-/**
- * Execute the bounded text-search action.
- *
- * Public `pattern` input is treated as a literal string by default. Callers can
- * opt into raw ripgrep regex semantics with `regex: true`; malformed regex
- * patterns are surfaced as explicit user-facing errors instead of being
- * collapsed into a misleading no-match response.
- */
-// biome-ignore lint/complexity/noExcessiveCognitiveComplexity: function has multiple distinct paths (validation, regex vs literal, summary vs detailed, structured vs text, zero vs results) that are clearer when explicit than when split
-export async function executePatternAction(
-  params: ActionParams,
-  cwd: string,
-): Promise<CodeIntelResult> {
-  if (!params.pattern) {
-    return {
-      content: "**Error:** `pattern` action requires a `pattern` parameter.",
-      details: undefined,
-    };
-  }
-
-  const maxResults = params.maxResults ?? 8;
-  const contextLines = params.contextLines ?? 1;
-  const scopePath = params.path ? normalizePath(params.path, cwd) : cwd;
-  const relScope = params.path ?? ".";
-
-  if (isStructuredPatternKind(params.kind)) {
-    const structured = await getStructuredPatternMatches(
-      { ...params, pattern: params.pattern, kind: params.kind },
-      scopePath,
-      cwd,
-      relScope,
-    );
-    if (typeof structured === "string") {
-      const errorDetails: SearchDetails = {
-        confidence: "unavailable",
-        scope: params.path ?? null,
-        candidateCount: 0,
-        omittedCount: 0,
-        nextQueries: ["Fix the regex pattern and retry"],
-      };
-      return { content: structured, details: { type: "search", data: errorDetails } };
-    }
-
-    if (structured) {
-      if (structured.matches.length === 0) {
-        return {
-          content: formatStructuredEmptyState(params.pattern, params.kind, relScope, structured),
-          details: {
-            type: "search",
-            data: {
-              confidence: "structural",
-              scope: params.path ?? null,
-              candidateCount: 0,
-              omittedCount: structured.omittedCount,
-              nextQueries: [
-                "Try a broader `pattern`, or omit `kind` for plain text search",
-                "Narrow `path` if the structured scan was partial",
-              ],
-            },
-          },
-        };
-      }
-
-      return {
-        content: formatStructuredMatches(params.pattern, params.kind, relScope, structured),
-        details: {
-          type: "search",
-          data: {
-            confidence: "structural",
-            scope: params.path ?? null,
-            candidateCount: structured.matches.length,
-            omittedCount: structured.omittedCount,
-            nextQueries: [
-              "Omit `kind` for plain text matches",
-              "Use `summary: true` for broader textual distribution",
-            ],
-          },
-        },
-      };
-    }
-  }
-
-  const matches = params.regex
-    ? getRegexMatches({
-        pattern: params.pattern,
-        scopePath,
-        cwd,
-        maxResults,
-        contextLines,
-        summary: params.summary,
-      })
-    : runRipgrep(escapeRegex(params.pattern), scopePath, cwd, {
-        maxMatches: params.summary ? undefined : maxResults * 3,
-        contextLines,
-        filterLowSignal: true,
-      });
-
-  if (typeof matches === "string") {
-    const errorDetails: SearchDetails = {
-      confidence: "unavailable",
-      scope: params.path ?? null,
-      candidateCount: 0,
-      omittedCount: 0,
-      nextQueries: ["Fix the regex pattern and retry"],
-    };
-    return { content: matches, details: { type: "search", data: errorDetails } };
-  }
-
-  if (matches.length === 0) {
-    const emptyDetails: SearchDetails = {
-      confidence: "heuristic",
-      scope: params.path ?? null,
-      candidateCount: 0,
-      omittedCount: 0,
-      nextQueries: params.regex
-        ? ["Set `regex: false` for literal matching"]
-        : ["Set `regex: true` for regex matching"],
-    };
-    return {
-      content: `No matches found for \`${params.pattern}\` in \`${relScope}\`.`,
-      details: { type: "search", data: emptyDetails },
-    };
-  }
-
-  const content = params.summary
-    ? formatPatternSummary(params.pattern, relScope, matches, maxResults)
-    : formatPatternResults(params.pattern, relScope, matches, maxResults);
-
-  const details: SearchDetails = {
-    confidence: "heuristic",
-    scope: params.path ?? null,
-    candidateCount: matches.length,
-    omittedCount: 0,
-    nextQueries: params.regex
-      ? ["Set `regex: false` for literal matching"]
-      : ["Set `regex: true` for regex matching"],
-  };
-  return { content, details: { type: "search" as const, data: details } };
-}
-
-function formatStructuredEmptyState(
-  pattern: string,
-  kind: "definition" | "export" | "import",
-  relScope: string,
-  result: StructuredPatternResult,
-): string {
-  const lines = [`No ${kind} matches found for \`${pattern}\` in \`${relScope}\`.`];
-  const partialWarning = formatPartialStructuredWarning(result);
-  if (partialWarning) {
-    lines.push("");
-    lines.push(partialWarning);
-  }
-  return lines.join("\n");
-}
-
-function formatStructuredMatches(
-  pattern: string,
-  kind: "definition" | "export" | "import",
-  relScope: string,
-  result: StructuredPatternResult,
-): string {
-  const grouped = new Map<string, StructuredMatch[]>();
-  for (const match of result.matches) {
-    const group = grouped.get(match.file) ?? [];
-    group.push(match);
-    grouped.set(match.file, group);
-  }
-
-  const kindLabel =
-    kind === "definition" ? "Definitions" : kind === "export" ? "Exports" : "Imports";
-  const lines: string[] = [];
-  lines.push(`# Pattern ${kindLabel}: \`${pattern}\``);
-  lines.push("");
-  lines.push(
-    `**${result.matches.length} match${result.matches.length !== 1 ? "es" : ""}** across **${grouped.size} file${grouped.size !== 1 ? "s" : ""}** in \`${relScope}\``,
-  );
-  const partialWarning = formatPartialStructuredWarning(result);
-  if (partialWarning) {
-    lines.push(partialWarning);
-  }
-  lines.push("");
-
-  if (kind === "definition" || kind === "export") {
-    addDuplicateSummary(lines, result.matches);
-  }
-
-  for (const [file, fileMatches] of grouped) {
-    lines.push(`### ${file}`);
-    for (const match of fileMatches.slice(0, 8)) {
-      lines.push(`- \`${match.name}\` (${match.kind}) L${match.line}`);
-    }
-    if (fileMatches.length > 8) {
-      lines.push(`- _+${fileMatches.length - 8} more in this file_`);
-    }
-    lines.push("");
-  }
-
-  return lines.join("\n");
-}
-
-function formatPartialStructuredWarning(result: StructuredPatternResult): string | null {
-  if (!result.partialReason || result.omittedCount <= 0) return null;
-
-  if (result.partialReason === "timeout") {
-    return `_Partial structured results — scan timed out with +${result.omittedCount} file${result.omittedCount !== 1 ? "s" : ""} omitted. Narrow \`path\` or \`pattern\` for complete coverage._`;
-  }
-
-  return `_Partial structured results — +${result.omittedCount} file${result.omittedCount !== 1 ? "s" : ""} omitted after reaching the structured scan cap. Narrow \`path\` or \`pattern\` for complete coverage._`;
-}
-
-function addDuplicateSummary(lines: string[], matches: StructuredMatch[]): void {
-  const byName = new Map<string, Set<string>>();
-  for (const match of matches) {
-    const files = byName.get(match.name) ?? new Set<string>();
-    files.add(match.file);
-    byName.set(match.name, files);
-  }
-
-  const duplicates = [...byName.entries()]
-    .map(([name, files]) => ({ name, files: [...files].sort((a, b) => a.localeCompare(b)) }))
-    .filter((entry) => entry.files.length > 1)
-    .sort((a, b) => b.files.length - a.files.length || a.name.localeCompare(b.name));
-
-  if (duplicates.length === 0) return;
-
-  lines.push("## Duplicate Definitions");
-  for (const duplicate of duplicates.slice(0, 8)) {
-    lines.push(
-      `- \`${duplicate.name}\` — defined in ${duplicate.files.length} files: ${duplicate.files
-        .map((file) => `\`${file}\``)
-        .join(", ")}`,
-    );
-  }
-  if (duplicates.length > 8) {
-    lines.push(`- _+${duplicates.length - 8} more duplicates_`);
-  }
-  lines.push("");
-}
-
-function getRegexMatches(options: {
-  pattern: string;
-  scopePath: string;
-  cwd: string;
-  maxResults: number;
-  contextLines: number;
-  summary?: boolean;
-}): RgMatch[] | string {
-  const result = runRipgrepDetailed(options.pattern, options.scopePath, options.cwd, {
-    maxMatches: options.summary ? undefined : options.maxResults * 3,
-    contextLines: options.contextLines,
-    filterLowSignal: true,
-  });
-
-  if (result.error) {
-    return formatRegexError(options.pattern, result.error);
-  }
-
-  return result.matches;
-}
-
-function formatRegexError(pattern: string, error: string): string {
-  const lines = error
-    .split("\n")
-    .map((line) => line.trim())
-    .filter(Boolean);
-  const detail =
-    [...lines]
-      .reverse()
-      .find((line) => line.startsWith("error:"))
-      ?.replace(/^error:\s*/, "") ??
-    lines.at(-1) ??
-    "ripgrep rejected the regex.";
-  return `**Error:** Invalid regex pattern \`${pattern}\`: ${detail}`;
-}
-
-function formatPatternResults(
-  pattern: string,
-  relScope: string,
-  matches: RgMatch[],
-  maxResults: number,
-): string {
-  const lines: string[] = [];
-  lines.push(`# Pattern: \`${pattern}\``);
-  lines.push("");
-  lines.push(`**${matches.length} match${matches.length > 1 ? "es" : ""}** in \`${relScope}\``);
-  lines.push("");
-
-  const byFile = groupByFile(matches);
-  let shown = 0;
-  for (const [file, fileMatches] of byFile) {
-    if (shown >= maxResults) break;
-    lines.push(`### ${file}`);
-    const renderedLines = new Set<number>();
-    const matchLines = new Set(fileMatches.map((match) => match.line));
-    for (const m of fileMatches.slice(0, 5)) {
-      renderMatchWithContext(lines, m, renderedLines, matchLines);
-    }
-    if (fileMatches.length > 5) {
-      lines.push(`- _+${fileMatches.length - 5} more in this file_`);
-    }
-    lines.push("");
-    shown++;
-  }
-
-  if (byFile.size > maxResults) {
-    lines.push(
-      `_+${byFile.size - maxResults} more files omitted. Narrow with \`path\` or increase \`maxResults\`._`,
-    );
-  }
-
-  return lines.join("\n");
-}
-
-function formatPatternSummary(
-  pattern: string,
-  relScope: string,
-  matches: RgMatch[],
-  maxResults: number,
-): string {
-  const byFile = groupByFile(matches);
-  const byDir = new Map<string, number>();
-  for (const [file] of byFile) {
-    const dir = file.includes("/") ? file.split("/").slice(0, -1).join("/") : ".";
-    byDir.set(dir, (byDir.get(dir) ?? 0) + 1);
-  }
-
-  const lines: string[] = [];
-  lines.push(`# Pattern Summary: \`${pattern}\``);
-  lines.push("");
-  lines.push(
-    `**${matches.length} match${matches.length > 1 ? "es" : ""}** across **${byFile.size} file${byFile.size !== 1 ? "s" : ""}** in \`${relScope}\``,
-  );
-  lines.push("");
-
-  const sortedDirs = [...byDir.entries()].sort((a, b) => b[1] - a[1]).slice(0, maxResults);
-  for (const [dir, count] of sortedDirs) {
-    lines.push(`- \`${dir}/\` — ${count} file${count !== 1 ? "s" : ""}`);
-  }
-
-  if (byDir.size > maxResults) {
-    lines.push(`- _+${byDir.size - maxResults} more directories_`);
-  }
-
-  lines.push("");
-  return lines.join("\n");
-}
-
-function renderMatchWithContext(
-  lines: string[],
-  m: RgMatch,
-  renderedLines: Set<number>,
-  matchLines: Set<number>,
-): void {
-  if (m.context) {
-    for (const c of m.context.filter((cx) => cx.line < m.line)) {
-      if (renderedLines.has(c.line) || matchLines.has(c.line)) continue;
-      lines.push(`  L${c.line}: ${c.text.slice(0, 120)}`);
-      renderedLines.add(c.line);
-    }
-  }
-  if (!renderedLines.has(m.line)) {
-    lines.push(`- L${m.line}: \`${m.text.slice(0, 120)}\``);
-    renderedLines.add(m.line);
-  }
-  if (m.context) {
-    for (const c of m.context.filter((cx) => cx.line > m.line)) {
-      if (renderedLines.has(c.line) || matchLines.has(c.line)) continue;
-      lines.push(`  L${c.line}: ${c.text.slice(0, 120)}`);
-      renderedLines.add(c.line);
-    }
-  }
-}

package/src/providers/semantic-provider.ts

@@ -1,34 +0,0 @@
-import {
-  getSessionLspService,
-  type SessionLspService,
-  type SessionLspServiceState,
-  waitForSessionLspService,
-} from "@mrclrchtr/supi-lsp/api";
-
-export interface SemanticProviderOptions {
-  waitForReady?: boolean;
-  timeoutMs?: number;
-}
-
-export const DEFAULT_SEMANTIC_WAIT_MS = 250;
-
-/** Acquire the current session-scoped LSP service, optionally waiting for startup. */
-export async function getSemanticService(
-  cwd: string,
-  options: SemanticProviderOptions = {},
-): Promise<SessionLspService | null> {
-  const state = await getSemanticServiceState(cwd, options);
-  return state.kind === "ready" ? state.service : null;
-}
-
-/** Read the current LSP service state with an optional short wait for readiness. */
-export async function getSemanticServiceState(
-  cwd: string,
-  options: SemanticProviderOptions = {},
-): Promise<SessionLspServiceState> {
-  const current = getSessionLspService(cwd);
-  if (!options.waitForReady || current.kind !== "pending") {
-    return current;
-  }
-  return waitForSessionLspService(cwd, options.timeoutMs ?? DEFAULT_SEMANTIC_WAIT_MS);
-}

package/src/providers/structural-provider.ts

@@ -1,26 +0,0 @@
-import {
-  createTreeSitterSession,
-  getSessionTreeSitterService,
-  type TreeSitterService,
-} from "@mrclrchtr/supi-tree-sitter/api";
-
-/**
- * Run work against the shared session-scoped Tree-sitter service when available,
- * falling back to a short-lived owned session otherwise.
- */
-export async function withStructuralSession<T>(
-  cwd: string,
-  fn: (session: TreeSitterService) => Promise<T>,
-): Promise<T> {
-  const current = getSessionTreeSitterService(cwd);
-  if (current.kind === "ready") {
-    return fn(current.service);
-  }
-
-  const session = createTreeSitterSession(cwd);
-  try {
-    return await fn(session);
-  } finally {
-    session.dispose();
-  }
-}