@tachu/extensions 1.0.0-alpha.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +119 -0
- package/LICENSE +201 -0
- package/README.md +1104 -0
- package/README_ZH.md +1082 -0
- package/dist/backends/file.d.ts +18 -0
- package/dist/backends/file.d.ts.map +1 -0
- package/dist/backends/file.js +85 -0
- package/dist/backends/file.js.map +1 -0
- package/dist/backends/index.d.ts +4 -0
- package/dist/backends/index.d.ts.map +1 -0
- package/dist/backends/index.js +4 -0
- package/dist/backends/index.js.map +1 -0
- package/dist/backends/terminal.d.ts +18 -0
- package/dist/backends/terminal.d.ts.map +1 -0
- package/dist/backends/terminal.js +81 -0
- package/dist/backends/terminal.js.map +1 -0
- package/dist/backends/web.d.ts +18 -0
- package/dist/backends/web.d.ts.map +1 -0
- package/dist/backends/web.js +55 -0
- package/dist/backends/web.js.map +1 -0
- package/dist/common/net.d.ts +39 -0
- package/dist/common/net.d.ts.map +1 -0
- package/dist/common/net.js +177 -0
- package/dist/common/net.js.map +1 -0
- package/dist/common/path.d.ts +51 -0
- package/dist/common/path.d.ts.map +1 -0
- package/dist/common/path.js +76 -0
- package/dist/common/path.js.map +1 -0
- package/dist/common/process.d.ts +19 -0
- package/dist/common/process.d.ts.map +1 -0
- package/dist/common/process.js +67 -0
- package/dist/common/process.js.map +1 -0
- package/dist/index.d.ts +13 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +13 -0
- package/dist/index.js.map +1 -0
- package/dist/mcp/index.d.ts +3 -0
- package/dist/mcp/index.d.ts.map +1 -0
- package/dist/mcp/index.js +3 -0
- package/dist/mcp/index.js.map +1 -0
- package/dist/mcp/sse-adapter.d.ts +82 -0
- package/dist/mcp/sse-adapter.d.ts.map +1 -0
- package/dist/mcp/sse-adapter.js +201 -0
- package/dist/mcp/sse-adapter.js.map +1 -0
- package/dist/mcp/stdio-adapter.d.ts +85 -0
- package/dist/mcp/stdio-adapter.d.ts.map +1 -0
- package/dist/mcp/stdio-adapter.js +203 -0
- package/dist/mcp/stdio-adapter.js.map +1 -0
- package/dist/memory/fs-memory-system.d.ts +147 -0
- package/dist/memory/fs-memory-system.d.ts.map +1 -0
- package/dist/memory/fs-memory-system.js +266 -0
- package/dist/memory/fs-memory-system.js.map +1 -0
- package/dist/memory/index.d.ts +2 -0
- package/dist/memory/index.d.ts.map +1 -0
- package/dist/memory/index.js +2 -0
- package/dist/memory/index.js.map +1 -0
- package/dist/observability/index.d.ts +3 -0
- package/dist/observability/index.d.ts.map +1 -0
- package/dist/observability/index.js +3 -0
- package/dist/observability/index.js.map +1 -0
- package/dist/observability/jsonl-emitter.d.ts +58 -0
- package/dist/observability/jsonl-emitter.d.ts.map +1 -0
- package/dist/observability/jsonl-emitter.js +96 -0
- package/dist/observability/jsonl-emitter.js.map +1 -0
- package/dist/observability/otel-emitter.d.ts +52 -0
- package/dist/observability/otel-emitter.d.ts.map +1 -0
- package/dist/observability/otel-emitter.js +143 -0
- package/dist/observability/otel-emitter.js.map +1 -0
- package/dist/providers/anthropic.d.ts +73 -0
- package/dist/providers/anthropic.d.ts.map +1 -0
- package/dist/providers/anthropic.js +521 -0
- package/dist/providers/anthropic.js.map +1 -0
- package/dist/providers/index.d.ts +5 -0
- package/dist/providers/index.d.ts.map +1 -0
- package/dist/providers/index.js +5 -0
- package/dist/providers/index.js.map +1 -0
- package/dist/providers/mock.d.ts +81 -0
- package/dist/providers/mock.d.ts.map +1 -0
- package/dist/providers/mock.js +160 -0
- package/dist/providers/mock.js.map +1 -0
- package/dist/providers/openai.d.ts +95 -0
- package/dist/providers/openai.d.ts.map +1 -0
- package/dist/providers/openai.js +529 -0
- package/dist/providers/openai.js.map +1 -0
- package/dist/providers/qwen.d.ts +145 -0
- package/dist/providers/qwen.d.ts.map +1 -0
- package/dist/providers/qwen.js +669 -0
- package/dist/providers/qwen.js.map +1 -0
- package/dist/rules/index.d.ts +9 -0
- package/dist/rules/index.d.ts.map +1 -0
- package/dist/rules/index.js +15 -0
- package/dist/rules/index.js.map +1 -0
- package/dist/rules/no-hallucination.md +11 -0
- package/dist/rules/no-sensitive-output.md +11 -0
- package/dist/rules/prefer-concise-response.md +11 -0
- package/dist/rules/require-tool-verification.md +11 -0
- package/dist/safety/default-gate.d.ts +112 -0
- package/dist/safety/default-gate.d.ts.map +1 -0
- package/dist/safety/default-gate.js +188 -0
- package/dist/safety/default-gate.js.map +1 -0
- package/dist/safety/index.d.ts +2 -0
- package/dist/safety/index.d.ts.map +1 -0
- package/dist/safety/index.js +2 -0
- package/dist/safety/index.js.map +1 -0
- package/dist/tools/_shared/web-client.d.ts +18 -0
- package/dist/tools/_shared/web-client.d.ts.map +1 -0
- package/dist/tools/_shared/web-client.js +46 -0
- package/dist/tools/_shared/web-client.js.map +1 -0
- package/dist/tools/apply-patch/descriptor.md +27 -0
- package/dist/tools/apply-patch/executor.d.ts +19 -0
- package/dist/tools/apply-patch/executor.d.ts.map +1 -0
- package/dist/tools/apply-patch/executor.js +193 -0
- package/dist/tools/apply-patch/executor.js.map +1 -0
- package/dist/tools/fetch-url/descriptor.md +44 -0
- package/dist/tools/fetch-url/executor.d.ts +28 -0
- package/dist/tools/fetch-url/executor.d.ts.map +1 -0
- package/dist/tools/fetch-url/executor.js +115 -0
- package/dist/tools/fetch-url/executor.js.map +1 -0
- package/dist/tools/index.d.ts +12 -0
- package/dist/tools/index.d.ts.map +1 -0
- package/dist/tools/index.js +286 -0
- package/dist/tools/index.js.map +1 -0
- package/dist/tools/list-dir/descriptor.md +29 -0
- package/dist/tools/list-dir/executor.d.ts +22 -0
- package/dist/tools/list-dir/executor.d.ts.map +1 -0
- package/dist/tools/list-dir/executor.js +48 -0
- package/dist/tools/list-dir/executor.js.map +1 -0
- package/dist/tools/read-file/descriptor.md +28 -0
- package/dist/tools/read-file/executor.d.ts +15 -0
- package/dist/tools/read-file/executor.d.ts.map +1 -0
- package/dist/tools/read-file/executor.js +22 -0
- package/dist/tools/read-file/executor.js.map +1 -0
- package/dist/tools/run-shell/descriptor.md +39 -0
- package/dist/tools/run-shell/executor.d.ts +20 -0
- package/dist/tools/run-shell/executor.d.ts.map +1 -0
- package/dist/tools/run-shell/executor.js +76 -0
- package/dist/tools/run-shell/executor.js.map +1 -0
- package/dist/tools/search-code/descriptor.md +31 -0
- package/dist/tools/search-code/executor.d.ts +23 -0
- package/dist/tools/search-code/executor.d.ts.map +1 -0
- package/dist/tools/search-code/executor.js +122 -0
- package/dist/tools/search-code/executor.js.map +1 -0
- package/dist/tools/shared.d.ts +47 -0
- package/dist/tools/shared.d.ts.map +1 -0
- package/dist/tools/shared.js +27 -0
- package/dist/tools/shared.js.map +1 -0
- package/dist/tools/web-fetch/descriptor.md +198 -0
- package/dist/tools/web-fetch/errors.d.ts +32 -0
- package/dist/tools/web-fetch/errors.d.ts.map +1 -0
- package/dist/tools/web-fetch/errors.js +91 -0
- package/dist/tools/web-fetch/errors.js.map +1 -0
- package/dist/tools/web-fetch/executor.d.ts +10 -0
- package/dist/tools/web-fetch/executor.d.ts.map +1 -0
- package/dist/tools/web-fetch/executor.js +191 -0
- package/dist/tools/web-fetch/executor.js.map +1 -0
- package/dist/tools/web-fetch/index.d.ts +4 -0
- package/dist/tools/web-fetch/index.d.ts.map +1 -0
- package/dist/tools/web-fetch/index.js +3 -0
- package/dist/tools/web-fetch/index.js.map +1 -0
- package/dist/tools/web-fetch/types.d.ts +157 -0
- package/dist/tools/web-fetch/types.d.ts.map +1 -0
- package/dist/tools/web-fetch/types.js +7 -0
- package/dist/tools/web-fetch/types.js.map +1 -0
- package/dist/tools/web-search/descriptor.md +89 -0
- package/dist/tools/web-search/errors.d.ts +33 -0
- package/dist/tools/web-search/errors.d.ts.map +1 -0
- package/dist/tools/web-search/errors.js +45 -0
- package/dist/tools/web-search/errors.js.map +1 -0
- package/dist/tools/web-search/executor.d.ts +10 -0
- package/dist/tools/web-search/executor.d.ts.map +1 -0
- package/dist/tools/web-search/executor.js +185 -0
- package/dist/tools/web-search/executor.js.map +1 -0
- package/dist/tools/web-search/index.d.ts +4 -0
- package/dist/tools/web-search/index.d.ts.map +1 -0
- package/dist/tools/web-search/index.js +3 -0
- package/dist/tools/web-search/index.js.map +1 -0
- package/dist/tools/web-search/types.d.ts +86 -0
- package/dist/tools/web-search/types.d.ts.map +1 -0
- package/dist/tools/web-search/types.js +7 -0
- package/dist/tools/web-search/types.js.map +1 -0
- package/dist/tools/write-file/descriptor.md +30 -0
- package/dist/tools/write-file/executor.d.ts +16 -0
- package/dist/tools/write-file/executor.d.ts.map +1 -0
- package/dist/tools/write-file/executor.js +18 -0
- package/dist/tools/write-file/executor.js.map +1 -0
- package/dist/transformers/document-to-text.d.ts +23 -0
- package/dist/transformers/document-to-text.d.ts.map +1 -0
- package/dist/transformers/document-to-text.js +69 -0
- package/dist/transformers/document-to-text.js.map +1 -0
- package/dist/transformers/image-to-text.d.ts +38 -0
- package/dist/transformers/image-to-text.d.ts.map +1 -0
- package/dist/transformers/image-to-text.js +82 -0
- package/dist/transformers/image-to-text.js.map +1 -0
- package/dist/transformers/index.d.ts +3 -0
- package/dist/transformers/index.d.ts.map +1 -0
- package/dist/transformers/index.js +3 -0
- package/dist/transformers/index.js.map +1 -0
- package/dist/vector/index.d.ts +3 -0
- package/dist/vector/index.d.ts.map +1 -0
- package/dist/vector/index.js +3 -0
- package/dist/vector/index.js.map +1 -0
- package/dist/vector/local-fs.d.ts +76 -0
- package/dist/vector/local-fs.d.ts.map +1 -0
- package/dist/vector/local-fs.js +153 -0
- package/dist/vector/local-fs.js.map +1 -0
- package/dist/vector/qdrant.d.ts +85 -0
- package/dist/vector/qdrant.d.ts.map +1 -0
- package/dist/vector/qdrant.js +208 -0
- package/dist/vector/qdrant.js.map +1 -0
- package/package.json +74 -0
|
@@ -0,0 +1,193 @@
|
|
|
1
|
+
import { mkdir, readFile, writeFile, rm } from "node:fs/promises";
|
|
2
|
+
import { dirname } from "node:path";
|
|
3
|
+
import { ValidationError } from "@tachu/core";
|
|
4
|
+
import { assertNotAborted, resolveSandboxPolicy } from "../shared";
|
|
5
|
+
import { resolveAllowedPath, toWorkspaceRelativePath } from "../../common/path";
|
|
6
|
+
const normalizePatchPath = (value) => value.replace(/^[ab]\//, "").trim();
|
|
7
|
+
const splitContent = (content) => {
|
|
8
|
+
if (content.length === 0) {
|
|
9
|
+
return { lines: [], trailingNewline: false };
|
|
10
|
+
}
|
|
11
|
+
const normalized = content.replaceAll("\r\n", "\n");
|
|
12
|
+
const trailingNewline = normalized.endsWith("\n");
|
|
13
|
+
const lines = normalized.split("\n");
|
|
14
|
+
if (trailingNewline) {
|
|
15
|
+
lines.pop();
|
|
16
|
+
}
|
|
17
|
+
return { lines, trailingNewline };
|
|
18
|
+
};
|
|
19
|
+
const parsePatch = (rawPatch) => {
|
|
20
|
+
const lines = rawPatch.replaceAll("\r\n", "\n").split("\n");
|
|
21
|
+
const files = [];
|
|
22
|
+
let index = 0;
|
|
23
|
+
while (index < lines.length) {
|
|
24
|
+
const line = lines[index] ?? "";
|
|
25
|
+
if (!line.startsWith("--- ")) {
|
|
26
|
+
index += 1;
|
|
27
|
+
continue;
|
|
28
|
+
}
|
|
29
|
+
const oldPath = normalizePatchPath(line.slice(4));
|
|
30
|
+
const newLine = lines[index + 1] ?? "";
|
|
31
|
+
if (!newLine.startsWith("+++ ")) {
|
|
32
|
+
throw new ValidationError("VALIDATION_PATCH_FORMAT", "patch 缺少 +++ 文件头");
|
|
33
|
+
}
|
|
34
|
+
const newPath = normalizePatchPath(newLine.slice(4));
|
|
35
|
+
index += 2;
|
|
36
|
+
const hunks = [];
|
|
37
|
+
while (index < lines.length) {
|
|
38
|
+
const hunkHeader = lines[index] ?? "";
|
|
39
|
+
if (hunkHeader.startsWith("--- ")) {
|
|
40
|
+
break;
|
|
41
|
+
}
|
|
42
|
+
if (!hunkHeader.startsWith("@@ ")) {
|
|
43
|
+
index += 1;
|
|
44
|
+
continue;
|
|
45
|
+
}
|
|
46
|
+
const match = /^@@ -(\d+)(?:,\d+)? \+(\d+)(?:,\d+)? @@/.exec(hunkHeader);
|
|
47
|
+
if (!match) {
|
|
48
|
+
throw new ValidationError("VALIDATION_PATCH_FORMAT", `无效 hunk 头: ${hunkHeader}`);
|
|
49
|
+
}
|
|
50
|
+
const oldStart = Number(match[1]);
|
|
51
|
+
const newStart = Number(match[2]);
|
|
52
|
+
index += 1;
|
|
53
|
+
const hunkLines = [];
|
|
54
|
+
while (index < lines.length) {
|
|
55
|
+
const hunkLine = lines[index] ?? "";
|
|
56
|
+
if (hunkLine.startsWith("@@ ") || hunkLine.startsWith("--- ")) {
|
|
57
|
+
break;
|
|
58
|
+
}
|
|
59
|
+
if (hunkLine === "\") {
|
|
60
|
+
index += 1;
|
|
61
|
+
continue;
|
|
62
|
+
}
|
|
63
|
+
const marker = hunkLine[0];
|
|
64
|
+
if (marker !== " " && marker !== "+" && marker !== "-") {
|
|
65
|
+
throw new ValidationError("VALIDATION_PATCH_FORMAT", `非法 hunk 行: ${hunkLine}`);
|
|
66
|
+
}
|
|
67
|
+
hunkLines.push({ type: marker, content: hunkLine.slice(1) });
|
|
68
|
+
index += 1;
|
|
69
|
+
}
|
|
70
|
+
hunks.push({ oldStart, newStart, lines: hunkLines });
|
|
71
|
+
}
|
|
72
|
+
files.push({ oldPath, newPath, hunks });
|
|
73
|
+
}
|
|
74
|
+
if (files.length === 0) {
|
|
75
|
+
throw new ValidationError("VALIDATION_PATCH_EMPTY", "patch 不包含任何文件");
|
|
76
|
+
}
|
|
77
|
+
return files;
|
|
78
|
+
};
|
|
79
|
+
const applyFilePatch = (original, patch) => {
|
|
80
|
+
const { lines: sourceLines, trailingNewline } = splitContent(original);
|
|
81
|
+
const result = [];
|
|
82
|
+
let cursor = 0;
|
|
83
|
+
for (const hunk of patch.hunks) {
|
|
84
|
+
const hunkStart = Math.max(0, hunk.oldStart - 1);
|
|
85
|
+
if (hunkStart < cursor) {
|
|
86
|
+
throw new ValidationError("VALIDATION_PATCH_CONFLICT", "patch hunk 重叠,无法应用");
|
|
87
|
+
}
|
|
88
|
+
result.push(...sourceLines.slice(cursor, hunkStart));
|
|
89
|
+
let sourceIndex = hunkStart;
|
|
90
|
+
for (const line of hunk.lines) {
|
|
91
|
+
if (line.type === " ") {
|
|
92
|
+
const actual = sourceLines[sourceIndex];
|
|
93
|
+
if (actual !== line.content) {
|
|
94
|
+
throw new ValidationError("VALIDATION_PATCH_CONFLICT", `上下文不匹配: expected "${line.content}" got "${actual ?? "<eof>"}"`);
|
|
95
|
+
}
|
|
96
|
+
result.push(actual);
|
|
97
|
+
sourceIndex += 1;
|
|
98
|
+
continue;
|
|
99
|
+
}
|
|
100
|
+
if (line.type === "-") {
|
|
101
|
+
const actual = sourceLines[sourceIndex];
|
|
102
|
+
if (actual !== line.content) {
|
|
103
|
+
throw new ValidationError("VALIDATION_PATCH_CONFLICT", `删除行不匹配: expected "${line.content}" got "${actual ?? "<eof>"}"`);
|
|
104
|
+
}
|
|
105
|
+
sourceIndex += 1;
|
|
106
|
+
continue;
|
|
107
|
+
}
|
|
108
|
+
result.push(line.content);
|
|
109
|
+
}
|
|
110
|
+
cursor = sourceIndex;
|
|
111
|
+
}
|
|
112
|
+
result.push(...sourceLines.slice(cursor));
|
|
113
|
+
return { content: result.join("\n"), trailingNewline };
|
|
114
|
+
};
|
|
115
|
+
/**
|
|
116
|
+
* 应用 unified diff Tool 执行器。
|
|
117
|
+
*/
|
|
118
|
+
export const applyPatchExecutor = async (input, context) => {
|
|
119
|
+
assertNotAborted(context.abortSignal);
|
|
120
|
+
const baseRoot = resolveAllowedPath(input.basePath ?? ".", resolveSandboxPolicy(context));
|
|
121
|
+
const filePatches = parsePatch(input.patch);
|
|
122
|
+
const backups = new Map();
|
|
123
|
+
const applied = [];
|
|
124
|
+
const rollback = async () => {
|
|
125
|
+
for (const [path, snapshot] of backups.entries()) {
|
|
126
|
+
if (!snapshot.exists) {
|
|
127
|
+
await rm(path, { force: true }).catch(() => undefined);
|
|
128
|
+
continue;
|
|
129
|
+
}
|
|
130
|
+
await mkdir(dirname(path), { recursive: true });
|
|
131
|
+
const suffix = snapshot.trailingNewline ? "\n" : "";
|
|
132
|
+
await writeFile(path, `${snapshot.content}${suffix}`);
|
|
133
|
+
}
|
|
134
|
+
};
|
|
135
|
+
try {
|
|
136
|
+
for (const patch of filePatches) {
|
|
137
|
+
assertNotAborted(context.abortSignal);
|
|
138
|
+
const logicalTarget = patch.newPath === "/dev/null" ? patch.oldPath : patch.newPath;
|
|
139
|
+
// 这里的校验语义与外层沙箱正交:patch 里声明的每个文件路径都必须
|
|
140
|
+
// 相对 `baseRoot` 不越界,即使外层整体已经 sandboxWaived 也不能让
|
|
141
|
+
// 单个 patch 跳出用户授权的 basePath(否则一次审批会变成无限通道)。
|
|
142
|
+
const absoluteTarget = resolveAllowedPath(logicalTarget, { allowedRoots: [baseRoot] });
|
|
143
|
+
if (!backups.has(absoluteTarget)) {
|
|
144
|
+
const existing = await readFile(absoluteTarget, "utf8").catch(() => null);
|
|
145
|
+
if (existing === null) {
|
|
146
|
+
backups.set(absoluteTarget, { exists: false, content: "", trailingNewline: false });
|
|
147
|
+
}
|
|
148
|
+
else {
|
|
149
|
+
const split = splitContent(existing);
|
|
150
|
+
backups.set(absoluteTarget, {
|
|
151
|
+
exists: true,
|
|
152
|
+
content: split.lines.join("\n"),
|
|
153
|
+
trailingNewline: split.trailingNewline,
|
|
154
|
+
});
|
|
155
|
+
}
|
|
156
|
+
}
|
|
157
|
+
const original = backups.get(absoluteTarget)?.exists
|
|
158
|
+
? (await readFile(absoluteTarget, "utf8").catch(() => "") ?? "")
|
|
159
|
+
: "";
|
|
160
|
+
const next = applyFilePatch(original, patch);
|
|
161
|
+
if (patch.newPath === "/dev/null") {
|
|
162
|
+
await rm(absoluteTarget, { force: true });
|
|
163
|
+
}
|
|
164
|
+
else {
|
|
165
|
+
await mkdir(dirname(absoluteTarget), { recursive: true });
|
|
166
|
+
const suffix = next.trailingNewline ? "\n" : "";
|
|
167
|
+
await writeFile(absoluteTarget, `${next.content}${suffix}`);
|
|
168
|
+
}
|
|
169
|
+
applied.push({
|
|
170
|
+
file: toWorkspaceRelativePath(context.workspaceRoot, absoluteTarget),
|
|
171
|
+
status: "ok",
|
|
172
|
+
});
|
|
173
|
+
}
|
|
174
|
+
return { applied, success: true };
|
|
175
|
+
}
|
|
176
|
+
catch (error) {
|
|
177
|
+
await rollback();
|
|
178
|
+
const reason = error instanceof Error ? error.message : String(error);
|
|
179
|
+
if (applied.length === 0) {
|
|
180
|
+
applied.push({ file: "<none>", status: "failed", reason });
|
|
181
|
+
}
|
|
182
|
+
else {
|
|
183
|
+
const previous = applied[applied.length - 1];
|
|
184
|
+
applied[applied.length - 1] = {
|
|
185
|
+
file: previous?.file ?? "<unknown>",
|
|
186
|
+
status: "failed",
|
|
187
|
+
reason,
|
|
188
|
+
};
|
|
189
|
+
}
|
|
190
|
+
return { applied, success: false };
|
|
191
|
+
}
|
|
192
|
+
};
|
|
193
|
+
//# sourceMappingURL=executor.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"executor.js","sourceRoot":"","sources":["../../../src/tools/apply-patch/executor.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAE,QAAQ,EAAE,SAAS,EAAE,EAAE,EAAE,MAAM,kBAAkB,CAAC;AAClE,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AACpC,OAAO,EAAE,eAAe,EAAE,MAAM,aAAa,CAAC;AAE9C,OAAO,EAAE,gBAAgB,EAAE,oBAAoB,EAAE,MAAM,WAAW,CAAC;AACnE,OAAO,EAAE,kBAAkB,EAAE,uBAAuB,EAAE,MAAM,mBAAmB,CAAC;AAmChF,MAAM,kBAAkB,GAAG,CAAC,KAAa,EAAU,EAAE,CACnD,KAAK,CAAC,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;AAEtC,MAAM,YAAY,GAAG,CAAC,OAAe,EAAiD,EAAE;IACtF,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACzB,OAAO,EAAE,KAAK,EAAE,EAAE,EAAE,eAAe,EAAE,KAAK,EAAE,CAAC;IAC/C,CAAC;IACD,MAAM,UAAU,GAAG,OAAO,CAAC,UAAU,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;IACpD,MAAM,eAAe,GAAG,UAAU,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;IAClD,MAAM,KAAK,GAAG,UAAU,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACrC,IAAI,eAAe,EAAE,CAAC;QACpB,KAAK,CAAC,GAAG,EAAE,CAAC;IACd,CAAC;IACD,OAAO,EAAE,KAAK,EAAE,eAAe,EAAE,CAAC;AACpC,CAAC,CAAC;AAEF,MAAM,UAAU,GAAG,CAAC,QAAgB,EAAe,EAAE;IACnD,MAAM,KAAK,GAAG,QAAQ,CAAC,UAAU,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC5D,MAAM,KAAK,GAAgB,EAAE,CAAC;IAC9B,IAAI,KAAK,GAAG,CAAC,CAAC;IAEd,OAAO,KAAK,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;QAC5B,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;QAChC,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,EAAE,CAAC;YAC7B,KAAK,IAAI,CAAC,CAAC;YACX,SAAS;QACX,CAAC;QACD,MAAM,OAAO,GAAG,kBAAkB,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QAClD,MAAM,OAAO,GAAG,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;QACvC,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,EAAE,CAAC;YAChC,MAAM,IAAI,eAAe,CAAC,yBAAyB,EAAE,kBAAkB,CAAC,CAAC;QAC3E,CAAC;QACD,MAAM,OAAO,GAAG,kBAAkB,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QACrD,KAAK,IAAI,CAAC,CAAC;QAEX,MAAM,KAAK,GAAW,EAAE,CAAC;QACzB,OAAO,KAAK,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;YAC5B,MAAM,UAAU,GAAG,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;YACtC,IAAI,UAAU,CAAC,UAAU,CAAC,MAAM,CAAC,EAAE,CAAC;gBAClC,MAAM;YACR,CAAC;YACD,IAAI,CAAC,UAAU,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;gBAClC,KAAK,IAAI,CAAC,CAAC;gBACX,SAAS;YACX,CAAC;YACD,MAAM,KAAK,GAAG,yCAAyC,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;YACzE,IAAI,CAAC,KAAK,EAAE,CAAC;gBACX,MAAM,IAAI,eAAe,CAAC,yBAAyB,EAAE,cAAc,UAAU,EAAE,CAAC,CAAC;YACnF,CAAC;YACD,MAAM,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;YAClC,MAAM,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;YAClC,KAAK,IAAI,CAAC,CAAC;YAEX,MAAM,SAAS,GAAe,EAAE,CAAC;YACjC,OAAO,KAAK,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;gBAC5B,MAAM,QAAQ,GAAG,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;gBACpC,IAAI,QAAQ,CAAC,UAAU,CAAC,KAAK,CAAC,IAAI,QAAQ,CAAC,UAAU,CAAC,MAAM,CAAC,EAAE,CAAC;oBAC9D,MAAM;gBACR,CAAC;gBACD,IAAI,QAAQ,KAAK,8BAA8B,EAAE,CAAC;oBAChD,KAAK,IAAI,CAAC,CAAC;oBACX,SAAS;gBACX,CAAC;gBACD,MAAM,MAAM,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;gBAC3B,IAAI,MAAM,KAAK,GAAG,IAAI,MAAM,KAAK,GAAG,IAAI,MAAM,KAAK,GAAG,EAAE,CAAC;oBACvD,MAAM,IAAI,eAAe,CAAC,yBAAyB,EAAE,cAAc,QAAQ,EAAE,CAAC,CAAC;gBACjF,CAAC;gBACD,SAAS,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;gBAC7D,KAAK,IAAI,CAAC,CAAC;YACb,CAAC;YACD,KAAK,CAAC,IAAI,CAAC,EAAE,QAAQ,EAAE,QAAQ,EAAE,KAAK,EAAE,SAAS,EAAE,CAAC,CAAC;QACvD,CAAC;QAED,KAAK,CAAC,IAAI,CAAC,EAAE,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC,CAAC;IAC1C,CAAC;IAED,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,MAAM,IAAI,eAAe,CAAC,wBAAwB,EAAE,eAAe,CAAC,CAAC;IACvE,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC,CAAC;AAEF,MAAM,cAAc,GAAG,CACrB,QAAgB,EAChB,KAAgB,EAC+B,EAAE;IACjD,MAAM,EAAE,KAAK,EAAE,WAAW,EAAE,eAAe,EAAE,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;IACvE,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,IAAI,MAAM,GAAG,CAAC,CAAC;IAEf,KAAK,MAAM,IAAI,IAAI,KAAK,CAAC,KAAK,EAAE,CAAC;QAC/B,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,QAAQ,GAAG,CAAC,CAAC,CAAC;QACjD,IAAI,SAAS,GAAG,MAAM,EAAE,CAAC;YACvB,MAAM,IAAI,eAAe,CAAC,2BAA2B,EAAE,oBAAoB,CAAC,CAAC;QAC/E,CAAC;QACD,MAAM,CAAC,IAAI,CAAC,GAAG,WAAW,CAAC,KAAK,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC,CAAC;QACrD,IAAI,WAAW,GAAG,SAAS,CAAC;QAE5B,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;YAC9B,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,EAAE,CAAC;gBACtB,MAAM,MAAM,GAAG,WAAW,CAAC,WAAW,CAAC,CAAC;gBACxC,IAAI,MAAM,KAAK,IAAI,CAAC,OAAO,EAAE,CAAC;oBAC5B,MAAM,IAAI,eAAe,CACvB,2BAA2B,EAC3B,qBAAqB,IAAI,CAAC,OAAO,UAAU,MAAM,IAAI,OAAO,GAAG,CAChE,CAAC;gBACJ,CAAC;gBACD,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;gBACpB,WAAW,IAAI,CAAC,CAAC;gBACjB,SAAS;YACX,CAAC;YACD,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,EAAE,CAAC;gBACtB,MAAM,MAAM,GAAG,WAAW,CAAC,WAAW,CAAC,CAAC;gBACxC,IAAI,MAAM,KAAK,IAAI,CAAC,OAAO,EAAE,CAAC;oBAC5B,MAAM,IAAI,eAAe,CACvB,2BAA2B,EAC3B,qBAAqB,IAAI,CAAC,OAAO,UAAU,MAAM,IAAI,OAAO,GAAG,CAChE,CAAC;gBACJ,CAAC;gBACD,WAAW,IAAI,CAAC,CAAC;gBACjB,SAAS;YACX,CAAC;YACD,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QAC5B,CAAC;QACD,MAAM,GAAG,WAAW,CAAC;IACvB,CAAC;IACD,MAAM,CAAC,IAAI,CAAC,GAAG,WAAW,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC;IAC1C,OAAO,EAAE,OAAO,EAAE,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,eAAe,EAAE,CAAC;AACzD,CAAC,CAAC;AAEF;;GAEG;AACH,MAAM,CAAC,MAAM,kBAAkB,GAAoD,KAAK,EACtF,KAAK,EACL,OAAO,EACP,EAAE;IACF,gBAAgB,CAAC,OAAO,CAAC,WAAW,CAAC,CAAC;IACtC,MAAM,QAAQ,GAAG,kBAAkB,CAAC,KAAK,CAAC,QAAQ,IAAI,GAAG,EAAE,oBAAoB,CAAC,OAAO,CAAC,CAAC,CAAC;IAC1F,MAAM,WAAW,GAAG,UAAU,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;IAC5C,MAAM,OAAO,GAAG,IAAI,GAAG,EAAwB,CAAC;IAChD,MAAM,OAAO,GAAsE,EAAE,CAAC;IAEtF,MAAM,QAAQ,GAAG,KAAK,IAAmB,EAAE;QACzC,KAAK,MAAM,CAAC,IAAI,EAAE,QAAQ,CAAC,IAAI,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC;YACjD,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,CAAC;gBACrB,MAAM,EAAE,CAAC,IAAI,EAAE,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,CAAC;gBACvD,SAAS;YACX,CAAC;YACD,MAAM,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;YAChD,MAAM,MAAM,GAAG,QAAQ,CAAC,eAAe,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC;YACpD,MAAM,SAAS,CAAC,IAAI,EAAE,GAAG,QAAQ,CAAC,OAAO,GAAG,MAAM,EAAE,CAAC,CAAC;QACxD,CAAC;IACH,CAAC,CAAC;IAEF,IAAI,CAAC;QACH,KAAK,MAAM,KAAK,IAAI,WAAW,EAAE,CAAC;YAChC,gBAAgB,CAAC,OAAO,CAAC,WAAW,CAAC,CAAC;YACtC,MAAM,aAAa,GAAG,KAAK,CAAC,OAAO,KAAK,WAAW,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC;YACpF,qCAAqC;YACrC,gDAAgD;YAChD,4CAA4C;YAC5C,MAAM,cAAc,GAAG,kBAAkB,CAAC,aAAa,EAAE,EAAE,YAAY,EAAE,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC;YACvF,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC,EAAE,CAAC;gBACjC,MAAM,QAAQ,GAAG,MAAM,QAAQ,CAAC,cAAc,EAAE,MAAM,CAAC,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,CAAC;gBAC1E,IAAI,QAAQ,KAAK,IAAI,EAAE,CAAC;oBACtB,OAAO,CAAC,GAAG,CAAC,cAAc,EAAE,EAAE,MAAM,EAAE,KAAK,EAAE,OAAO,EAAE,EAAE,EAAE,eAAe,EAAE,KAAK,EAAE,CAAC,CAAC;gBACtF,CAAC;qBAAM,CAAC;oBACN,MAAM,KAAK,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;oBACrC,OAAO,CAAC,GAAG,CAAC,cAAc,EAAE;wBAC1B,MAAM,EAAE,IAAI;wBACZ,OAAO,EAAE,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC;wBAC/B,eAAe,EAAE,KAAK,CAAC,eAAe;qBACvC,CAAC,CAAC;gBACL,CAAC;YACH,CAAC;YAED,MAAM,QAAQ,GAAG,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC,EAAE,MAAM;gBAClD,CAAC,CAAC,CAAE,MAAM,QAAQ,CAAC,cAAc,EAAE,MAAM,CAAC,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,EAAE,CAAY,IAAI,EAAE,CAAC;gBAC5E,CAAC,CAAC,EAAE,CAAC;YAEP,MAAM,IAAI,GAAG,cAAc,CAAC,QAAQ,EAAE,KAAK,CAAC,CAAC;YAC7C,IAAI,KAAK,CAAC,OAAO,KAAK,WAAW,EAAE,CAAC;gBAClC,MAAM,EAAE,CAAC,cAAc,EAAE,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC,CAAC;YAC5C,CAAC;iBAAM,CAAC;gBACN,MAAM,KAAK,CAAC,OAAO,CAAC,cAAc,CAAC,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;gBAC1D,MAAM,MAAM,GAAG,IAAI,CAAC,eAAe,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC;gBAChD,MAAM,SAAS,CAAC,cAAc,EAAE,GAAG,IAAI,CAAC,OAAO,GAAG,MAAM,EAAE,CAAC,CAAC;YAC9D,CAAC;YACD,OAAO,CAAC,IAAI,CAAC;gBACX,IAAI,EAAE,uBAAuB,CAAC,OAAO,CAAC,aAAa,EAAE,cAAc,CAAC;gBACpE,MAAM,EAAE,IAAI;aACb,CAAC,CAAC;QACL,CAAC;QACD,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;IACpC,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,MAAM,QAAQ,EAAE,CAAC;QACjB,MAAM,MAAM,GAAG,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QACtE,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACzB,OAAO,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,QAAQ,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,EAAE,CAAC,CAAC;QAC7D,CAAC;aAAM,CAAC;YACN,MAAM,QAAQ,GAAG,OAAO,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;YAC7C,OAAO,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,GAAG;gBAC5B,IAAI,EAAE,QAAQ,EAAE,IAAI,IAAI,WAAW;gBACnC,MAAM,EAAE,QAAQ;gBAChB,MAAM;aACP,CAAC;QACJ,CAAC;QACD,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC;IACrC,CAAC;AACH,CAAC,CAAC"}
|
|
@@ -0,0 +1,44 @@
|
|
|
1
|
+
---
|
|
2
|
+
kind: tool
|
|
3
|
+
name: fetch-url
|
|
4
|
+
description: 发送 HTTP 请求并返回响应内容
|
|
5
|
+
sideEffect: readonly
|
|
6
|
+
idempotent: false
|
|
7
|
+
requiresApproval: false
|
|
8
|
+
timeout: 15000
|
|
9
|
+
inputSchema:
|
|
10
|
+
type: object
|
|
11
|
+
properties:
|
|
12
|
+
url:
|
|
13
|
+
type: string
|
|
14
|
+
method:
|
|
15
|
+
type: string
|
|
16
|
+
enum: [GET, POST]
|
|
17
|
+
headers:
|
|
18
|
+
type: object
|
|
19
|
+
body:
|
|
20
|
+
type: string
|
|
21
|
+
timeoutMs:
|
|
22
|
+
type: number
|
|
23
|
+
required: [url]
|
|
24
|
+
outputSchema:
|
|
25
|
+
type: object
|
|
26
|
+
properties:
|
|
27
|
+
status:
|
|
28
|
+
type: number
|
|
29
|
+
headers:
|
|
30
|
+
type: object
|
|
31
|
+
body:
|
|
32
|
+
type: string
|
|
33
|
+
truncated:
|
|
34
|
+
type: boolean
|
|
35
|
+
contentType:
|
|
36
|
+
type: string
|
|
37
|
+
execute: fetch-url
|
|
38
|
+
---
|
|
39
|
+
|
|
40
|
+
执行受控网络请求,默认阻止私网地址并限制响应体大小。
|
|
41
|
+
|
|
42
|
+
响应体返回给 LLM 前会经过两道处理:
|
|
43
|
+
1. `Content-Type: text/html` / `application/xhtml+xml` 时,会剥掉 `<script>` / `<style>` / `<noscript>` / `<svg>` / `<canvas>` 及 HTML 注释,再合并空白,得到一段信噪比更高的正文;
|
|
44
|
+
2. 无论类型,都会按字符数上限(32KB 字符)截断;超出时在末尾追加 `... [内容已截断,完整长度 N 字符]` 提示。`truncated=true` 表示字节层或字符层任一发生了截断。
|
|
@@ -0,0 +1,28 @@
|
|
|
1
|
+
import type { ToolExecutor } from "../shared";
|
|
2
|
+
interface FetchUrlInput {
|
|
3
|
+
url: string;
|
|
4
|
+
method?: "GET" | "POST";
|
|
5
|
+
headers?: Record<string, string>;
|
|
6
|
+
body?: string;
|
|
7
|
+
timeoutMs?: number;
|
|
8
|
+
}
|
|
9
|
+
interface FetchUrlOutput {
|
|
10
|
+
status: number;
|
|
11
|
+
headers: Record<string, string>;
|
|
12
|
+
body: string;
|
|
13
|
+
truncated: boolean;
|
|
14
|
+
contentType?: string;
|
|
15
|
+
}
|
|
16
|
+
/**
|
|
17
|
+
* URL 抓取 Tool 执行器。
|
|
18
|
+
*
|
|
19
|
+
* 返回给 LLM 的 body 经过两道处理:
|
|
20
|
+
* 1. 若 `Content-Type: text/html` / `application/xhtml+xml` → HTML → 纯文本清洗
|
|
21
|
+
* 2. 无论什么类型 → 按字符数截断到 {@link MAX_BODY_CHARS},尾部追加截断提示
|
|
22
|
+
*
|
|
23
|
+
* `truncated` 为 true 时表示**字节层或字符层任一**发生了截断;`contentType` 归一化为小写主
|
|
24
|
+
* MIME(例如 `text/html` / `application/json`),供上层调用方做进一步判断。
|
|
25
|
+
*/
|
|
26
|
+
export declare const fetchUrlExecutor: ToolExecutor<FetchUrlInput, FetchUrlOutput>;
|
|
27
|
+
export {};
|
|
28
|
+
//# sourceMappingURL=executor.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"executor.d.ts","sourceRoot":"","sources":["../../../src/tools/fetch-url/executor.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,WAAW,CAAC;AAG9C,UAAU,aAAa;IACrB,GAAG,EAAE,MAAM,CAAC;IACZ,MAAM,CAAC,EAAE,KAAK,GAAG,MAAM,CAAC;IACxB,OAAO,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IACjC,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,SAAS,CAAC,EAAE,MAAM,CAAC;CACpB;AAED,UAAU,cAAc;IACtB,MAAM,EAAE,MAAM,CAAC;IACf,OAAO,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAChC,IAAI,EAAE,MAAM,CAAC;IACb,SAAS,EAAE,OAAO,CAAC;IACnB,WAAW,CAAC,EAAE,MAAM,CAAC;CACtB;AAqED;;;;;;;;;GASG;AACH,eAAO,MAAM,gBAAgB,EAAE,YAAY,CAAC,aAAa,EAAE,cAAc,CA4CxE,CAAC"}
|
|
@@ -0,0 +1,115 @@
|
|
|
1
|
+
import { assertPublicUrl, readResponseBodyWithLimit, withAbortTimeout } from "../../common/net";
|
|
2
|
+
import { assertNotAborted } from "../shared";
|
|
3
|
+
/**
|
|
4
|
+
* 响应体字节数上限(保护网络与内存层)。超过会走 {@link readResponseBodyWithLimit} 的尾部截断。
|
|
5
|
+
*/
|
|
6
|
+
const MAX_BODY_BYTES = 5 * 1024 * 1024;
|
|
7
|
+
/**
|
|
8
|
+
* 返回给 LLM 的字符数上限。经验值:32KB 字符 ≈ 8k~10k tokens,足以容纳一个文档页面的正文,
|
|
9
|
+
* 又不至于把下一轮 Agentic Loop 的 context 吹爆。超过会在末尾追加可见的截断提示。
|
|
10
|
+
*/
|
|
11
|
+
const MAX_BODY_CHARS = 32 * 1024;
|
|
12
|
+
/**
|
|
13
|
+
* 从 `Content-Type` 头提取主 MIME 类型(小写、去掉 `charset=` 等参数)。
|
|
14
|
+
*/
|
|
15
|
+
const parseContentType = (raw) => {
|
|
16
|
+
if (!raw)
|
|
17
|
+
return undefined;
|
|
18
|
+
const main = raw.split(";")[0]?.trim().toLowerCase();
|
|
19
|
+
return main && main.length > 0 ? main : undefined;
|
|
20
|
+
};
|
|
21
|
+
/**
|
|
22
|
+
* HTML → 纯文本的极简清洗(不引入第三方解析器):
|
|
23
|
+
* 1. 剥 `<!-- ... -->` 注释
|
|
24
|
+
* 2. 剥 `<script>` / `<style>` / `<noscript>` / `<svg>` / `<canvas>` 块(含内容)
|
|
25
|
+
* 3. 剥剩余标签
|
|
26
|
+
* 4. 解码常见 HTML 实体
|
|
27
|
+
* 5. 归一化空白(多空格 / 多空行合并)
|
|
28
|
+
*
|
|
29
|
+
* 目的不是 100% 还原原文结构,而是让 LLM 拿到一段信噪比尚可的正文。
|
|
30
|
+
*/
|
|
31
|
+
const stripHtmlForLlm = (html) => {
|
|
32
|
+
return html
|
|
33
|
+
.replace(/<!--[\s\S]*?-->/g, "")
|
|
34
|
+
.replace(/<script\b[^>]*>[\s\S]*?<\/script>/gi, "")
|
|
35
|
+
.replace(/<style\b[^>]*>[\s\S]*?<\/style>/gi, "")
|
|
36
|
+
.replace(/<noscript\b[^>]*>[\s\S]*?<\/noscript>/gi, "")
|
|
37
|
+
.replace(/<svg\b[^>]*>[\s\S]*?<\/svg>/gi, "")
|
|
38
|
+
.replace(/<canvas\b[^>]*>[\s\S]*?<\/canvas>/gi, "")
|
|
39
|
+
.replace(/<[^>]+>/g, " ")
|
|
40
|
+
.replace(/ /g, " ")
|
|
41
|
+
.replace(/&/g, "&")
|
|
42
|
+
.replace(/</g, "<")
|
|
43
|
+
.replace(/>/g, ">")
|
|
44
|
+
.replace(/"/g, '"')
|
|
45
|
+
.replace(/'/g, "'")
|
|
46
|
+
.replace(/[ \t]+/g, " ")
|
|
47
|
+
.replace(/\n{3,}/g, "\n\n")
|
|
48
|
+
.trim();
|
|
49
|
+
};
|
|
50
|
+
/**
|
|
51
|
+
* 按字符数截断,超出时在末尾追加可见的截断提示(让 LLM 知道自己只拿到部分内容)。
|
|
52
|
+
*/
|
|
53
|
+
const clipToChars = (text) => {
|
|
54
|
+
if (text.length <= MAX_BODY_CHARS)
|
|
55
|
+
return { text, truncated: false };
|
|
56
|
+
const clipped = text.slice(0, MAX_BODY_CHARS);
|
|
57
|
+
return {
|
|
58
|
+
text: `${clipped}\n\n... [内容已截断,完整长度 ${text.length} 字符]`,
|
|
59
|
+
truncated: true,
|
|
60
|
+
};
|
|
61
|
+
};
|
|
62
|
+
const isHtmlLike = (contentType) => {
|
|
63
|
+
if (!contentType)
|
|
64
|
+
return false;
|
|
65
|
+
return contentType === "text/html" || contentType === "application/xhtml+xml";
|
|
66
|
+
};
|
|
67
|
+
/**
|
|
68
|
+
* URL 抓取 Tool 执行器。
|
|
69
|
+
*
|
|
70
|
+
* 返回给 LLM 的 body 经过两道处理:
|
|
71
|
+
* 1. 若 `Content-Type: text/html` / `application/xhtml+xml` → HTML → 纯文本清洗
|
|
72
|
+
* 2. 无论什么类型 → 按字符数截断到 {@link MAX_BODY_CHARS},尾部追加截断提示
|
|
73
|
+
*
|
|
74
|
+
* `truncated` 为 true 时表示**字节层或字符层任一**发生了截断;`contentType` 归一化为小写主
|
|
75
|
+
* MIME(例如 `text/html` / `application/json`),供上层调用方做进一步判断。
|
|
76
|
+
*/
|
|
77
|
+
export const fetchUrlExecutor = async (input, context) => {
|
|
78
|
+
assertNotAborted(context.abortSignal);
|
|
79
|
+
const parsed = await assertPublicUrl(input.url);
|
|
80
|
+
const timeout = withAbortTimeout(context.abortSignal, input.timeoutMs ?? 15_000, "TIMEOUT_FETCH_URL");
|
|
81
|
+
try {
|
|
82
|
+
const response = await fetch(parsed, {
|
|
83
|
+
method: input.method ?? "GET",
|
|
84
|
+
headers: input.headers,
|
|
85
|
+
body: input.body,
|
|
86
|
+
signal: timeout.signal,
|
|
87
|
+
});
|
|
88
|
+
const bodyResult = await readResponseBodyWithLimit(response, MAX_BODY_BYTES);
|
|
89
|
+
const headers = {};
|
|
90
|
+
response.headers.forEach((value, key) => {
|
|
91
|
+
headers[key] = value;
|
|
92
|
+
});
|
|
93
|
+
const contentType = parseContentType(response.headers.get("content-type"));
|
|
94
|
+
let body = bodyResult.body;
|
|
95
|
+
let truncated = bodyResult.truncated;
|
|
96
|
+
if (isHtmlLike(contentType)) {
|
|
97
|
+
body = stripHtmlForLlm(body);
|
|
98
|
+
}
|
|
99
|
+
const clipped = clipToChars(body);
|
|
100
|
+
body = clipped.text;
|
|
101
|
+
if (clipped.truncated)
|
|
102
|
+
truncated = true;
|
|
103
|
+
return {
|
|
104
|
+
status: response.status,
|
|
105
|
+
headers,
|
|
106
|
+
body,
|
|
107
|
+
truncated,
|
|
108
|
+
...(contentType ? { contentType } : {}),
|
|
109
|
+
};
|
|
110
|
+
}
|
|
111
|
+
finally {
|
|
112
|
+
timeout.cleanup();
|
|
113
|
+
}
|
|
114
|
+
};
|
|
115
|
+
//# sourceMappingURL=executor.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"executor.js","sourceRoot":"","sources":["../../../src/tools/fetch-url/executor.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,eAAe,EAAE,yBAAyB,EAAE,gBAAgB,EAAE,MAAM,kBAAkB,CAAC;AAEhG,OAAO,EAAE,gBAAgB,EAAE,MAAM,WAAW,CAAC;AAkB7C;;GAEG;AACH,MAAM,cAAc,GAAG,CAAC,GAAG,IAAI,GAAG,IAAI,CAAC;AAEvC;;;GAGG;AACH,MAAM,cAAc,GAAG,EAAE,GAAG,IAAI,CAAC;AAEjC;;GAEG;AACH,MAAM,gBAAgB,GAAG,CAAC,GAAkB,EAAsB,EAAE;IAClE,IAAI,CAAC,GAAG;QAAE,OAAO,SAAS,CAAC;IAC3B,MAAM,IAAI,GAAG,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IACrD,OAAO,IAAI,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC;AACpD,CAAC,CAAC;AAEF;;;;;;;;;GASG;AACH,MAAM,eAAe,GAAG,CAAC,IAAY,EAAU,EAAE;IAC/C,OAAO,IAAI;SACR,OAAO,CAAC,kBAAkB,EAAE,EAAE,CAAC;SAC/B,OAAO,CAAC,qCAAqC,EAAE,EAAE,CAAC;SAClD,OAAO,CAAC,mCAAmC,EAAE,EAAE,CAAC;SAChD,OAAO,CAAC,yCAAyC,EAAE,EAAE,CAAC;SACtD,OAAO,CAAC,+BAA+B,EAAE,EAAE,CAAC;SAC5C,OAAO,CAAC,qCAAqC,EAAE,EAAE,CAAC;SAClD,OAAO,CAAC,UAAU,EAAE,GAAG,CAAC;SACxB,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC;SACvB,OAAO,CAAC,QAAQ,EAAE,GAAG,CAAC;SACtB,OAAO,CAAC,OAAO,EAAE,GAAG,CAAC;SACrB,OAAO,CAAC,OAAO,EAAE,GAAG,CAAC;SACrB,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC;SACvB,OAAO,CAAC,QAAQ,EAAE,GAAG,CAAC;SACtB,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC;SACvB,OAAO,CAAC,SAAS,EAAE,MAAM,CAAC;SAC1B,IAAI,EAAE,CAAC;AACZ,CAAC,CAAC;AAEF;;GAEG;AACH,MAAM,WAAW,GAAG,CAAC,IAAY,EAAwC,EAAE;IACzE,IAAI,IAAI,CAAC,MAAM,IAAI,cAAc;QAAE,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC;IACrE,MAAM,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,cAAc,CAAC,CAAC;IAC9C,OAAO;QACL,IAAI,EAAE,GAAG,OAAO,uBAAuB,IAAI,CAAC,MAAM,MAAM;QACxD,SAAS,EAAE,IAAI;KAChB,CAAC;AACJ,CAAC,CAAC;AAEF,MAAM,UAAU,GAAG,CAAC,WAAoB,EAAW,EAAE;IACnD,IAAI,CAAC,WAAW;QAAE,OAAO,KAAK,CAAC;IAC/B,OAAO,WAAW,KAAK,WAAW,IAAI,WAAW,KAAK,uBAAuB,CAAC;AAChF,CAAC,CAAC;AAEF;;;;;;;;;GASG;AACH,MAAM,CAAC,MAAM,gBAAgB,GAAgD,KAAK,EAChF,KAAK,EACL,OAAO,EACP,EAAE;IACF,gBAAgB,CAAC,OAAO,CAAC,WAAW,CAAC,CAAC;IACtC,MAAM,MAAM,GAAG,MAAM,eAAe,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IAChD,MAAM,OAAO,GAAG,gBAAgB,CAC9B,OAAO,CAAC,WAAW,EACnB,KAAK,CAAC,SAAS,IAAI,MAAM,EACzB,mBAAmB,CACpB,CAAC;IACF,IAAI,CAAC;QACH,MAAM,QAAQ,GAAG,MAAM,KAAK,CAAC,MAAM,EAAE;YACnC,MAAM,EAAE,KAAK,CAAC,MAAM,IAAI,KAAK;YAC7B,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,IAAI,EAAE,KAAK,CAAC,IAAI;YAChB,MAAM,EAAE,OAAO,CAAC,MAAM;SACvB,CAAC,CAAC;QACH,MAAM,UAAU,GAAG,MAAM,yBAAyB,CAAC,QAAQ,EAAE,cAAc,CAAC,CAAC;QAC7E,MAAM,OAAO,GAA2B,EAAE,CAAC;QAC3C,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,GAAG,EAAE,EAAE;YACtC,OAAO,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC;QACvB,CAAC,CAAC,CAAC;QACH,MAAM,WAAW,GAAG,gBAAgB,CAAC,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC,CAAC,CAAC;QAE3E,IAAI,IAAI,GAAG,UAAU,CAAC,IAAI,CAAC;QAC3B,IAAI,SAAS,GAAG,UAAU,CAAC,SAAS,CAAC;QACrC,IAAI,UAAU,CAAC,WAAW,CAAC,EAAE,CAAC;YAC5B,IAAI,GAAG,eAAe,CAAC,IAAI,CAAC,CAAC;QAC/B,CAAC;QACD,MAAM,OAAO,GAAG,WAAW,CAAC,IAAI,CAAC,CAAC;QAClC,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC;QACpB,IAAI,OAAO,CAAC,SAAS;YAAE,SAAS,GAAG,IAAI,CAAC;QAExC,OAAO;YACL,MAAM,EAAE,QAAQ,CAAC,MAAM;YACvB,OAAO;YACP,IAAI;YACJ,SAAS;YACT,GAAG,CAAC,WAAW,CAAC,CAAC,CAAC,EAAE,WAAW,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;SACxC,CAAC;IACJ,CAAC;YAAS,CAAC;QACT,OAAO,CAAC,OAAO,EAAE,CAAC;IACpB,CAAC;AACH,CAAC,CAAC"}
|
|
@@ -0,0 +1,12 @@
|
|
|
1
|
+
import type { ToolDescriptor } from "@tachu/core";
|
|
2
|
+
import type { ToolExecutor } from "./shared";
|
|
3
|
+
/**
|
|
4
|
+
* 内置工具描述符列表。
|
|
5
|
+
*/
|
|
6
|
+
export declare const toolDescriptors: ToolDescriptor[];
|
|
7
|
+
/**
|
|
8
|
+
* 工具执行函数注册表。
|
|
9
|
+
*/
|
|
10
|
+
export declare const toolExecutors: Record<string, ToolExecutor>;
|
|
11
|
+
export type { ToolExecutor, ToolExecutionContext } from "./shared";
|
|
12
|
+
//# sourceMappingURL=index.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/tools/index.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,aAAa,CAAC;AAClD,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,UAAU,CAAC;AAW7C;;GAEG;AACH,eAAO,MAAM,eAAe,EAAE,cAAc,EAqQ3C,CAAC;AAEF;;GAEG;AACH,eAAO,MAAM,aAAa,EAAE,MAAM,CAAC,MAAM,EAAE,YAAY,CAUtD,CAAC;AAEF,YAAY,EAAE,YAAY,EAAE,oBAAoB,EAAE,MAAM,UAAU,CAAC"}
|