opencode-gbk-tools 0.1.0 → 0.1.2

package/dist/opencode-tools/gbk_read.js

@@ -16240,6 +16240,8 @@ tool.schema = external_exports;
 
 // src/lib/gbk-file.ts
 var import_iconv_lite = __toESM(require_lib(), 1);
+import crypto from "crypto";
+import { createReadStream } from "fs";
 import fs2 from "fs/promises";
 import path2 from "path";
 
@@ -16302,6 +16304,7 @@ async function assertPathAllowed(filePath, context, allowExternal = false) {
 }
 
 // src/lib/gbk-file.ts
+var STREAMING_FILE_SIZE_THRESHOLD_BYTES = 1024 * 1024;
 function assertEncodingSupported(encoding) {
   if (encoding !== "gbk" && encoding !== "gb18030") {
     throw createGbkError("GBK_INVALID_ENCODING", `\u4E0D\u652F\u6301\u7684\u7F16\u7801: ${encoding}`);
@@ -16340,15 +16343,56 @@ function splitLinesWithNumbers(text, offset = 1, limit = 2e3) {
     content
   };
 }
+function detectNewlineStyle(text) {
+  const crlfCount = (text.match(/\r\n/g) ?? []).length;
+  const lfCount = (text.match(/(^|[^\r])\n/g) ?? []).length;
+  if (crlfCount > 0 && lfCount === 0) {
+    return "crlf";
+  }
+  if (lfCount > 0 && crlfCount === 0) {
+    return "lf";
+  }
+  if (crlfCount > 0 && lfCount > 0) {
+    return "mixed";
+  }
+  return "none";
+}
+function collectTailLines(text, limit) {
+  assertPositiveInteger(limit, "limit");
+  const lines = text.split(/\r?\n/);
+  const totalLines = lines.length;
+  const startLine = Math.max(1, totalLines - limit + 1);
+  const selected = lines.slice(startLine - 1);
+  return {
+    startLine,
+    endLine: totalLines,
+    totalLines,
+    content: selected.map((line, index) => `${startLine + index}: ${line}`).join("\n"),
+    truncated: startLine > 1,
+    tail: true,
+    newlineStyle: detectNewlineStyle(text)
+  };
+}
+function finalizeNewlineStyle(crlfCount, lfCount) {
+  if (crlfCount > 0 && lfCount === 0) {
+    return "crlf";
+  }
+  if (lfCount > 0 && crlfCount === 0) {
+    return "lf";
+  }
+  if (crlfCount > 0 && lfCount > 0) {
+    return "mixed";
+  }
+  return "none";
+}
 async function readBufferAsText(buffer, encoding) {
   assertEncodingSupported(encoding);
   assertNotBinary(buffer);
   return import_iconv_lite.default.decode(buffer, encoding);
 }
-async function readGbkFile(input) {
+async function resolveReadableGbkFile(input) {
   const encoding = input.encoding ?? "gbk";
-  const offset = input.offset ?? 1;
-  const limit = input.limit ?? 2e3;
+  assertEncodingSupported(encoding);
   const { candidatePath } = await assertPathAllowed(input.filePath, input.context, input.allowExternal ?? false);
   let stat;
   try {
@@ -16359,21 +16403,197 @@ async function readGbkFile(input) {
   if (stat.isDirectory()) {
     throw createGbkError("GBK_IS_DIRECTORY", `\u76EE\u6807\u8DEF\u5F84\u662F\u76EE\u5F55: ${candidatePath}`);
   }
+  return {
+    filePath: candidatePath,
+    encoding,
+    stat
+  };
+}
+async function readWholeGbkTextFile(input) {
   try {
-    const buffer = await fs2.readFile(candidatePath);
-    const text = await readBufferAsText(buffer, encoding);
+    const buffer = await fs2.readFile(input.filePath);
+    const content = await readBufferAsText(buffer, input.encoding);
     return {
-      filePath: candidatePath,
-      encoding,
-      ...splitLinesWithNumbers(text, offset, limit)
+      filePath: input.filePath,
+      encoding: input.encoding,
+      content,
+      bytesRead: buffer.byteLength
     };
   } catch (error45) {
     if (error45 instanceof Error && "code" in error45) {
       throw error45;
     }
-    throw createGbkError("GBK_IO_ERROR", `\u8BFB\u53D6\u6587\u4EF6\u5931\u8D25: ${candidatePath}`, error45);
+    throw createGbkError("GBK_IO_ERROR", `\u8BFB\u53D6\u6587\u4EF6\u5931\u8D25: ${input.filePath}`, error45);
   }
 }
+async function visitDecodedTextChunks(input, visitor) {
+  const decoder = import_iconv_lite.default.getDecoder(input.encoding);
+  const stream = createReadStream(input.filePath);
+  try {
+    for await (const chunk of stream) {
+      const buffer = Buffer.isBuffer(chunk) ? chunk : Buffer.from(chunk);
+      assertNotBinary(buffer);
+      const text = decoder.write(buffer);
+      if (text.length > 0) {
+        await visitor(text);
+      }
+    }
+    const trailingText = decoder.end();
+    if (trailingText && trailingText.length > 0) {
+      await visitor(trailingText);
+    }
+  } catch (error45) {
+    if (error45 instanceof Error && "code" in error45) {
+      throw error45;
+    }
+    throw createGbkError("GBK_IO_ERROR", `\u8BFB\u53D6\u6587\u4EF6\u5931\u8D25: ${input.filePath}`, error45);
+  } finally {
+    stream.destroy();
+  }
+}
+function createLineCollector(offset, limit) {
+  const lines = [];
+  let pending = "";
+  let totalLines = 0;
+  let crlfCount = 0;
+  let lfCount = 0;
+  const emitLine = (line) => {
+    totalLines += 1;
+    if (totalLines >= offset && lines.length < limit) {
+      lines.push(`${totalLines}: ${line}`);
+    }
+  };
+  return {
+    push(text) {
+      if (text.length === 0) {
+        return;
+      }
+      const combined = pending + text;
+      let start = 0;
+      while (true) {
+        const newlineIndex = combined.indexOf("\n", start);
+        if (newlineIndex === -1) {
+          break;
+        }
+        let line = combined.slice(start, newlineIndex);
+        if (line.endsWith("\r")) {
+          line = line.slice(0, -1);
+          crlfCount += 1;
+        } else {
+          lfCount += 1;
+        }
+        emitLine(line);
+        start = newlineIndex + 1;
+      }
+      pending = combined.slice(start);
+    },
+    finish() {
+      emitLine(pending);
+      const endLine = lines.length === 0 ? totalLines : offset + lines.length - 1;
+      return {
+        startLine: offset,
+        endLine,
+        totalLines,
+        content: lines.join("\n"),
+        tail: false,
+        truncated: endLine < totalLines,
+        newlineStyle: finalizeNewlineStyle(crlfCount, lfCount)
+      };
+    }
+  };
+}
+function createTailCollector(limit) {
+  assertPositiveInteger(limit, "limit");
+  const lines = [];
+  let pending = "";
+  let totalLines = 0;
+  let crlfCount = 0;
+  let lfCount = 0;
+  const emitLine = (line) => {
+    totalLines += 1;
+    lines.push(line);
+    if (lines.length > limit) {
+      lines.shift();
+    }
+  };
+  return {
+    push(text) {
+      if (text.length === 0) {
+        return;
+      }
+      const combined = pending + text;
+      let start = 0;
+      while (true) {
+        const newlineIndex = combined.indexOf("\n", start);
+        if (newlineIndex === -1) {
+          break;
+        }
+        let line = combined.slice(start, newlineIndex);
+        if (line.endsWith("\r")) {
+          line = line.slice(0, -1);
+          crlfCount += 1;
+        } else {
+          lfCount += 1;
+        }
+        emitLine(line);
+        start = newlineIndex + 1;
+      }
+      pending = combined.slice(start);
+    },
+    finish() {
+      emitLine(pending);
+      const startLine = Math.max(1, totalLines - lines.length + 1);
+      return {
+        startLine,
+        endLine: totalLines,
+        totalLines,
+        content: lines.map((line, index) => `${startLine + index}: ${line}`).join("\n"),
+        truncated: startLine > 1,
+        tail: true,
+        newlineStyle: finalizeNewlineStyle(crlfCount, lfCount)
+      };
+    }
+  };
+}
+async function readGbkFile(input) {
+  const offset = input.offset ?? 1;
+  const limit = input.limit ?? 2e3;
+  const tail = input.tail ?? false;
+  const resolved = await resolveReadableGbkFile(input);
+  if (resolved.stat.size < STREAMING_FILE_SIZE_THRESHOLD_BYTES) {
+    const textFile = await readWholeGbkTextFile(resolved);
+    const lineWindow2 = tail ? collectTailLines(textFile.content, limit) : {
+      ...splitLinesWithNumbers(textFile.content, offset, limit),
+      tail: false,
+      truncated: false,
+      newlineStyle: detectNewlineStyle(textFile.content)
+    };
+    if (!tail) {
+      lineWindow2.truncated = lineWindow2.endLine < lineWindow2.totalLines;
+    }
+    return {
+      filePath: textFile.filePath,
+      encoding: textFile.encoding,
+      bytesRead: textFile.bytesRead,
+      fileSize: textFile.bytesRead,
+      streamed: false,
+      ...lineWindow2
+    };
+  }
+  const collector = tail ? createTailCollector(limit) : createLineCollector(offset, limit);
+  await visitDecodedTextChunks(resolved, (text) => {
+    collector.push(text);
+  });
+  const lineWindow = collector.finish();
+  return {
+    filePath: resolved.filePath,
+    encoding: resolved.encoding,
+    bytesRead: resolved.stat.size,
+    fileSize: resolved.stat.size,
+    streamed: true,
+    ...lineWindow
+  };
+}
 
 // src/tools/gbk_read.ts
 var gbk_read_default = tool({
@@ -16382,6 +16602,7 @@ var gbk_read_default = tool({
     filePath: tool.schema.string().describe("Target file path"),
     offset: tool.schema.number().int().positive().optional().describe("1-based start line"),
     limit: tool.schema.number().int().positive().optional().describe("Number of lines to read"),
+    tail: tool.schema.boolean().optional().describe("Read last N lines instead of offset-based window"),
     encoding: tool.schema.enum(["gbk", "gb18030"]).optional().describe("Text encoding"),
     allowExternal: tool.schema.boolean().optional().describe("Allow paths outside workspace root")
   },

package/dist/opencode-tools/gbk_write.js

@@ -16240,6 +16240,8 @@ tool.schema = external_exports;
 
 // src/lib/gbk-file.ts
 var import_iconv_lite = __toESM(require_lib(), 1);
+import crypto from "crypto";
+import { createReadStream } from "fs";
 import fs2 from "fs/promises";
 import path2 from "path";
 
@@ -16302,6 +16304,7 @@ async function assertPathAllowed(filePath, context, allowExternal = false) {
 }
 
 // src/lib/gbk-file.ts
+var STREAMING_FILE_SIZE_THRESHOLD_BYTES = 1024 * 1024;
 function assertEncodingSupported(encoding) {
   if (encoding !== "gbk" && encoding !== "gb18030") {
     throw createGbkError("GBK_INVALID_ENCODING", `\u4E0D\u652F\u6301\u7684\u7F16\u7801: ${encoding}`);