@markitdownjs/cli 0.1.1 → 0.1.3
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/dist/commands/batch.d.ts +1 -1
- package/dist/commands/batch.d.ts.map +1 -1
- package/dist/commands/batch.js +14 -14
- package/dist/commands/batch.js.map +1 -1
- package/dist/commands/convert.d.ts +1 -1
- package/dist/commands/convert.d.ts.map +1 -1
- package/dist/commands/convert.js +14 -14
- package/dist/commands/convert.js.map +1 -1
- package/dist/commands/formats.d.ts +1 -1
- package/dist/commands/formats.js +10 -10
- package/dist/commands/formats.js.map +1 -1
- package/dist/commands/serve.d.ts +1 -1
- package/dist/commands/serve.d.ts.map +1 -1
- package/dist/commands/serve.js +27 -27
- package/dist/commands/serve.js.map +1 -1
- package/dist/commands/watch.d.ts +1 -1
- package/dist/commands/watch.d.ts.map +1 -1
- package/dist/commands/watch.js +10 -10
- package/dist/commands/watch.js.map +1 -1
- package/dist/index.js +9 -9
- package/package.json +4 -4
package/dist/commands/batch.d.ts
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"batch.d.ts","sourceRoot":"","sources":["../../src/commands/batch.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAKpC,wBAAgB,oBAAoB,CAAC,OAAO,EAAE,OAAO,GAAG,IAAI,
|
|
1
|
+
{"version":3,"file":"batch.d.ts","sourceRoot":"","sources":["../../src/commands/batch.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAKpC,wBAAgB,oBAAoB,CAAC,OAAO,EAAE,OAAO,GAAG,IAAI,CAoF3D"}
|
package/dist/commands/batch.js
CHANGED
|
@@ -1,19 +1,19 @@
|
|
|
1
|
-
import { readFile, writeFile, readdir, mkdir } from
|
|
2
|
-
import { join, extname, resolve, relative, dirname } from
|
|
3
|
-
import { MarkItDown } from
|
|
1
|
+
import { readFile, writeFile, readdir, mkdir } from "fs/promises";
|
|
2
|
+
import { join, extname, resolve, relative, dirname } from "path";
|
|
3
|
+
import { MarkItDown } from "@markitdownjs/core";
|
|
4
4
|
export function registerBatchCommand(program) {
|
|
5
5
|
program
|
|
6
|
-
.command(
|
|
7
|
-
.description(
|
|
8
|
-
.argument(
|
|
9
|
-
.option(
|
|
10
|
-
.option(
|
|
11
|
-
.option(
|
|
12
|
-
.option(
|
|
6
|
+
.command("batch")
|
|
7
|
+
.description("Batch convert all files in a directory")
|
|
8
|
+
.argument("[dir]", "Directory to process", ".")
|
|
9
|
+
.option("-o, --output <dir>", "Output directory")
|
|
10
|
+
.option("-e, --extensions <exts>", "Comma-separated extensions to process", ".pdf,.docx,.xlsx,.pptx,.html,.csv,.json,.xml")
|
|
11
|
+
.option("-c, --concurrency <n>", "Max concurrent conversions", "5")
|
|
12
|
+
.option("-v, --verbose", "Verbose output")
|
|
13
13
|
.action(async (dir, options) => {
|
|
14
14
|
const inputDir = resolve(dir);
|
|
15
15
|
const outputDir = options.output ? resolve(options.output) : inputDir;
|
|
16
|
-
const extensions = options.extensions.split(
|
|
16
|
+
const extensions = options.extensions.split(",").map((e) => e.trim());
|
|
17
17
|
const concurrency = parseInt(options.concurrency, 10);
|
|
18
18
|
const parser = new MarkItDown();
|
|
19
19
|
async function getFiles(dirPath) {
|
|
@@ -22,7 +22,7 @@ export function registerBatchCommand(program) {
|
|
|
22
22
|
for (const entry of entries) {
|
|
23
23
|
const fullPath = join(dirPath, entry.name);
|
|
24
24
|
if (entry.isDirectory()) {
|
|
25
|
-
files.push(...await getFiles(fullPath));
|
|
25
|
+
files.push(...(await getFiles(fullPath)));
|
|
26
26
|
}
|
|
27
27
|
else if (entry.isFile() && extensions.includes(extname(entry.name).toLowerCase())) {
|
|
28
28
|
files.push(fullPath);
|
|
@@ -43,9 +43,9 @@ export function registerBatchCommand(program) {
|
|
|
43
43
|
fileName: filePath,
|
|
44
44
|
});
|
|
45
45
|
const relPath = relative(inputDir, filePath);
|
|
46
|
-
const mdPath = join(outputDir, relPath.replace(extname(filePath),
|
|
46
|
+
const mdPath = join(outputDir, relPath.replace(extname(filePath), ".md"));
|
|
47
47
|
await mkdir(dirname(mdPath), { recursive: true });
|
|
48
|
-
await writeFile(mdPath, result.markdown,
|
|
48
|
+
await writeFile(mdPath, result.markdown, "utf-8");
|
|
49
49
|
converted++;
|
|
50
50
|
if (options.verbose) {
|
|
51
51
|
console.error(`[${converted}/${files.length}] Converted: ${relPath}`);
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"batch.js","sourceRoot":"","sources":["../../src/commands/batch.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,QAAQ,EAAE,SAAS,EAAE,OAAO,EAAE,KAAK,EAAE,MAAM,aAAa,CAAC;AAClE,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,MAAM,MAAM,CAAC;AACjE,OAAO,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAEhD,MAAM,UAAU,oBAAoB,CAAC,OAAgB;IACnD,OAAO;SACJ,OAAO,CAAC,OAAO,CAAC;SAChB,WAAW,CAAC,wCAAwC,CAAC;SACrD,QAAQ,CAAC,OAAO,EAAE,sBAAsB,EAAE,GAAG,CAAC;SAC9C,MAAM,CAAC,oBAAoB,EAAE,kBAAkB,CAAC;SAChD,MAAM,
|
|
1
|
+
{"version":3,"file":"batch.js","sourceRoot":"","sources":["../../src/commands/batch.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,QAAQ,EAAE,SAAS,EAAE,OAAO,EAAE,KAAK,EAAE,MAAM,aAAa,CAAC;AAClE,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,MAAM,MAAM,CAAC;AACjE,OAAO,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAEhD,MAAM,UAAU,oBAAoB,CAAC,OAAgB;IACnD,OAAO;SACJ,OAAO,CAAC,OAAO,CAAC;SAChB,WAAW,CAAC,wCAAwC,CAAC;SACrD,QAAQ,CAAC,OAAO,EAAE,sBAAsB,EAAE,GAAG,CAAC;SAC9C,MAAM,CAAC,oBAAoB,EAAE,kBAAkB,CAAC;SAChD,MAAM,CACL,yBAAyB,EACzB,uCAAuC,EACvC,8CAA8C,CAC/C;SACA,MAAM,CAAC,uBAAuB,EAAE,4BAA4B,EAAE,GAAG,CAAC;SAClE,MAAM,CAAC,eAAe,EAAE,gBAAgB,CAAC;SACzC,MAAM,CACL,KAAK,EACH,GAAW,EACX,OAAuF,EACvF,EAAE;QACF,MAAM,QAAQ,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC;QAC9B,MAAM,SAAS,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC;QACtE,MAAM,UAAU,GAAG,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;QACtE,MAAM,WAAW,GAAG,QAAQ,CAAC,OAAO,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC;QACtD,MAAM,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;QAEhC,KAAK,UAAU,QAAQ,CAAC,OAAe;YACrC,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,OAAO,EAAE,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC,CAAC;YAChE,MAAM,KAAK,GAAa,EAAE,CAAC;YAE3B,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;gBAC5B,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC;gBAC3C,IAAI,KAAK,CAAC,WAAW,EAAE,EAAE,CAAC;oBACxB,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC;gBAC5C,CAAC;qBAAM,IAAI,KAAK,CAAC,MAAM,EAAE,IAAI,UAAU,CAAC,QAAQ,CAAC,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,WAAW,EAAE,CAAC,EAAE,CAAC;oBACpF,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;gBACvB,CAAC;YACH,CAAC;YAED,OAAO,KAAK,CAAC;QACf,CAAC;QAED,IAAI,CAAC;YACH,MAAM,KAAK,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAC;YACvC,IAAI,SAAS,GAAG,CAAC,CAAC;YAClB,IAAI,MAAM,GAAG,CAAC,CAAC;YAEf,OAAO,CAAC,KAAK,CAAC,SAAS,KAAK,CAAC,MAAM,mBAAmB,CAAC,CAAC;YAExD,KAAK,UAAU,WAAW,CAAC,QAAgB;gBACzC,IAAI,CAAC;oBACH,MAAM,OAAO,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAC;oBACzC,MAAM,MAAM,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC;wBAClC,IAAI,EAAE,IAAI,UAAU,CAAC,OAAO,CAAC;wBAC7B,QAAQ,EAAE,QAAQ;qBACnB,CAAC,CAAC;oBAEH,MAAM,OAAO,GAAG,QAAQ,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC;oBAC7C,MAAM,MAAM,GAAG,IAAI,CAAC,SAAS,EAAE,OAAO,CAAC,OAAO,CAAC,OAAO,CAAC,QAAQ,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC;oBAC1E,MAAM,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;oBAClD,MAAM,SAAS,CAAC,MAAM,EAAE,MAAM,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;oBAElD,SAAS,EAAE,CAAC;oBACZ,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;wBACpB,OAAO,CAAC,KAAK,CAAC,IAAI,SAAS,IAAI,KAAK,CAAC,MAAM,gBAAgB,OAAO,EAAE,CAAC,CAAC;oBACxE,CAAC;gBACH,CAAC;gBAAC,OAAO,KAAK,EAAE,CAAC;oBACf,MAAM,EAAE,CAAC;oBACT,OAAO,CAAC,KAAK,CACX,WAAW,QAAQ,MAAM,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAClF,CAAC;gBACJ,CAAC;YACH,CAAC;YAED,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,WAAW,EAAE,CAAC;gBACnD,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,WAAW,CAAC,CAAC;gBAC9C,MAAM,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC,CAAC;YAC5C,CAAC;YAED,OAAO,CAAC,KAAK,CAAC,qBAAqB,SAAS,eAAe,MAAM,SAAS,CAAC,CAAC;QAC9E,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,CAAC,KAAK,CAAC,UAAU,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;YAClF,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAClB,CAAC;IACH,CAAC,CACF,CAAC;AACN,CAAC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"convert.d.ts","sourceRoot":"","sources":["../../src/commands/convert.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAKpC,wBAAgB,sBAAsB,CAAC,OAAO,EAAE,OAAO,GAAG,IAAI,
|
|
1
|
+
{"version":3,"file":"convert.d.ts","sourceRoot":"","sources":["../../src/commands/convert.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAKpC,wBAAgB,sBAAsB,CAAC,OAAO,EAAE,OAAO,GAAG,IAAI,CA8D7D"}
|
package/dist/commands/convert.js
CHANGED
|
@@ -1,14 +1,14 @@
|
|
|
1
|
-
import { readFile, writeFile, mkdir } from
|
|
2
|
-
import { dirname } from
|
|
3
|
-
import { MarkItDown } from
|
|
1
|
+
import { readFile, writeFile, mkdir } from "fs/promises";
|
|
2
|
+
import { dirname } from "path";
|
|
3
|
+
import { MarkItDown } from "@markitdownjs/core";
|
|
4
4
|
export function registerConvertCommand(program) {
|
|
5
5
|
program
|
|
6
|
-
.command(
|
|
7
|
-
.description(
|
|
8
|
-
.argument(
|
|
9
|
-
.option(
|
|
10
|
-
.option(
|
|
11
|
-
.option(
|
|
6
|
+
.command("convert")
|
|
7
|
+
.description("Convert a document to Markdown")
|
|
8
|
+
.argument("<file>", "File to convert")
|
|
9
|
+
.option("-o, --output <path>", "Output file path")
|
|
10
|
+
.option("-f, --format <format>", "Output format (markdown, json, plaintext, html)", "markdown")
|
|
11
|
+
.option("-v, --verbose", "Verbose output")
|
|
12
12
|
.action(async (file, options) => {
|
|
13
13
|
try {
|
|
14
14
|
if (options.verbose) {
|
|
@@ -25,23 +25,23 @@ export function registerConvertCommand(program) {
|
|
|
25
25
|
});
|
|
26
26
|
let output;
|
|
27
27
|
switch (options.format) {
|
|
28
|
-
case
|
|
28
|
+
case "json":
|
|
29
29
|
output = JSON.stringify(result, null, 2);
|
|
30
30
|
break;
|
|
31
|
-
case
|
|
31
|
+
case "plaintext":
|
|
32
32
|
output = result.markdown;
|
|
33
33
|
break;
|
|
34
|
-
case
|
|
34
|
+
case "html":
|
|
35
35
|
output = result.markdown;
|
|
36
36
|
break;
|
|
37
|
-
case
|
|
37
|
+
case "markdown":
|
|
38
38
|
default:
|
|
39
39
|
output = result.markdown;
|
|
40
40
|
break;
|
|
41
41
|
}
|
|
42
42
|
if (options.output) {
|
|
43
43
|
await mkdir(dirname(options.output), { recursive: true });
|
|
44
|
-
await writeFile(options.output, output,
|
|
44
|
+
await writeFile(options.output, output, "utf-8");
|
|
45
45
|
if (options.verbose) {
|
|
46
46
|
console.error(`Written to: ${options.output}`);
|
|
47
47
|
}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"convert.js","sourceRoot":"","sources":["../../src/commands/convert.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,QAAQ,EAAE,SAAS,EAAE,KAAK,EAAE,MAAM,aAAa,CAAC;AACzD,OAAO,EAAE,OAAO,EAAE,MAAM,MAAM,CAAC;AAC/B,OAAO,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAEhD,MAAM,UAAU,sBAAsB,CAAC,OAAgB;IACrD,OAAO;SACJ,OAAO,CAAC,SAAS,CAAC;SAClB,WAAW,CAAC,gCAAgC,CAAC;SAC7C,QAAQ,CAAC,QAAQ,EAAE,iBAAiB,CAAC;SACrC,MAAM,CAAC,qBAAqB,EAAE,kBAAkB,CAAC;SACjD,MAAM,CAAC,uBAAuB,EAAE,iDAAiD,EAAE,UAAU,CAAC;SAC9F,MAAM,CAAC,eAAe,EAAE,gBAAgB,CAAC;SACzC,MAAM,
|
|
1
|
+
{"version":3,"file":"convert.js","sourceRoot":"","sources":["../../src/commands/convert.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,QAAQ,EAAE,SAAS,EAAE,KAAK,EAAE,MAAM,aAAa,CAAC;AACzD,OAAO,EAAE,OAAO,EAAE,MAAM,MAAM,CAAC;AAC/B,OAAO,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAEhD,MAAM,UAAU,sBAAsB,CAAC,OAAgB;IACrD,OAAO;SACJ,OAAO,CAAC,SAAS,CAAC;SAClB,WAAW,CAAC,gCAAgC,CAAC;SAC7C,QAAQ,CAAC,QAAQ,EAAE,iBAAiB,CAAC;SACrC,MAAM,CAAC,qBAAqB,EAAE,kBAAkB,CAAC;SACjD,MAAM,CAAC,uBAAuB,EAAE,iDAAiD,EAAE,UAAU,CAAC;SAC9F,MAAM,CAAC,eAAe,EAAE,gBAAgB,CAAC;SACzC,MAAM,CACL,KAAK,EAAE,IAAY,EAAE,OAA8D,EAAE,EAAE;QACrF,IAAI,CAAC;YACH,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;gBACpB,OAAO,CAAC,KAAK,CAAC,eAAe,IAAI,EAAE,CAAC,CAAC;YACvC,CAAC;YAED,MAAM,UAAU,GAAG,MAAM,QAAQ,CAAC,IAAI,CAAC,CAAC;YAExC,MAAM,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;YAChC,MAAM,MAAM,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC;gBAClC,IAAI,EAAE,IAAI,UAAU,CAAC,UAAU,CAAC;gBAChC,QAAQ,EAAE,IAAI;gBACd,OAAO,EAAE;oBACP,YAAY,EAAE,OAAO,CAAC,MAAoD;iBAC3E;aACF,CAAC,CAAC;YAEH,IAAI,MAAc,CAAC;YACnB,QAAQ,OAAO,CAAC,MAAM,EAAE,CAAC;gBACvB,KAAK,MAAM;oBACT,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;oBACzC,MAAM;gBACR,KAAK,WAAW;oBACd,MAAM,GAAG,MAAM,CAAC,QAAQ,CAAC;oBACzB,MAAM;gBACR,KAAK,MAAM;oBACT,MAAM,GAAG,MAAM,CAAC,QAAQ,CAAC;oBACzB,MAAM;gBACR,KAAK,UAAU,CAAC;gBAChB;oBACE,MAAM,GAAG,MAAM,CAAC,QAAQ,CAAC;oBACzB,MAAM;YACV,CAAC;YAED,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;gBACnB,MAAM,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;gBAC1D,MAAM,SAAS,CAAC,OAAO,CAAC,MAAM,EAAE,MAAM,EAAE,OAAO,CAAC,CAAC;gBACjD,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;oBACpB,OAAO,CAAC,KAAK,CAAC,eAAe,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC;gBACjD,CAAC;YACH,CAAC;iBAAM,CAAC;gBACN,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;YAC/B,CAAC;YAED,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;gBACpB,OAAO,CAAC,KAAK,CAAC,2BAA2B,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;YACjF,CAAC;QACH,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,CAAC,KAAK,CAAC,UAAU,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;YAClF,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAClB,CAAC;IACH,CAAC,CACF,CAAC;AACN,CAAC"}
|
package/dist/commands/formats.js
CHANGED
|
@@ -1,22 +1,22 @@
|
|
|
1
|
-
import { getSupportedExtensions, getSupportedMimeTypes } from
|
|
1
|
+
import { getSupportedExtensions, getSupportedMimeTypes } from "@markitdownjs/shared";
|
|
2
2
|
export function registerFormatsCommand(program) {
|
|
3
3
|
program
|
|
4
|
-
.command(
|
|
5
|
-
.description(
|
|
4
|
+
.command("formats")
|
|
5
|
+
.description("List all supported file formats")
|
|
6
6
|
.action(() => {
|
|
7
7
|
const extensions = getSupportedExtensions();
|
|
8
8
|
const mimeTypes = getSupportedMimeTypes();
|
|
9
|
-
console.log(
|
|
10
|
-
console.log(
|
|
9
|
+
console.log("Supported Formats:");
|
|
10
|
+
console.log("==================\n");
|
|
11
11
|
const rows = [];
|
|
12
12
|
for (let i = 0; i < extensions.length; i++) {
|
|
13
|
-
rows.push([extensions[i] ??
|
|
13
|
+
rows.push([extensions[i] ?? "", mimeTypes[i] ?? ""]);
|
|
14
14
|
}
|
|
15
|
-
const maxExt = Math.max(...rows.map(r => (r[0] ??
|
|
16
|
-
console.log(`${
|
|
17
|
-
console.log(`${
|
|
15
|
+
const maxExt = Math.max(...rows.map((r) => (r[0] ?? "").length), "Extension".length);
|
|
16
|
+
console.log(`${"Extension".padEnd(maxExt)} MIME Type`);
|
|
17
|
+
console.log(`${"-".repeat(maxExt)} ${"-".repeat(40)}`);
|
|
18
18
|
for (const row of rows) {
|
|
19
|
-
console.log(`${(row[0] ??
|
|
19
|
+
console.log(`${(row[0] ?? "").padEnd(maxExt)} ${row[1] ?? ""}`);
|
|
20
20
|
}
|
|
21
21
|
console.log(`\nTotal: ${extensions.length} formats supported`);
|
|
22
22
|
});
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"formats.js","sourceRoot":"","sources":["../../src/commands/formats.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,sBAAsB,EAAE,qBAAqB,EAAE,MAAM,sBAAsB,CAAC;AAErF,MAAM,UAAU,sBAAsB,CAAC,OAAgB;IACrD,OAAO;SACJ,OAAO,CAAC,SAAS,CAAC;SAClB,WAAW,CAAC,iCAAiC,CAAC;SAC9C,MAAM,CAAC,GAAG,EAAE;QACX,MAAM,UAAU,GAAG,sBAAsB,EAAE,CAAC;QAC5C,MAAM,SAAS,GAAG,qBAAqB,EAAE,CAAC;QAE1C,OAAO,CAAC,GAAG,CAAC,oBAAoB,CAAC,CAAC;QAClC,OAAO,CAAC,GAAG,CAAC,sBAAsB,CAAC,CAAC;QAEpC,MAAM,IAAI,GAAe,EAAE,CAAC;QAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC3C,IAAI,CAAC,IAAI,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,SAAS,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;QACvD,CAAC;QAED,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,EAAE,WAAW,CAAC,MAAM,CAAC,CAAC;
|
|
1
|
+
{"version":3,"file":"formats.js","sourceRoot":"","sources":["../../src/commands/formats.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,sBAAsB,EAAE,qBAAqB,EAAE,MAAM,sBAAsB,CAAC;AAErF,MAAM,UAAU,sBAAsB,CAAC,OAAgB;IACrD,OAAO;SACJ,OAAO,CAAC,SAAS,CAAC;SAClB,WAAW,CAAC,iCAAiC,CAAC;SAC9C,MAAM,CAAC,GAAG,EAAE;QACX,MAAM,UAAU,GAAG,sBAAsB,EAAE,CAAC;QAC5C,MAAM,SAAS,GAAG,qBAAqB,EAAE,CAAC;QAE1C,OAAO,CAAC,GAAG,CAAC,oBAAoB,CAAC,CAAC;QAClC,OAAO,CAAC,GAAG,CAAC,sBAAsB,CAAC,CAAC;QAEpC,MAAM,IAAI,GAAe,EAAE,CAAC;QAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC3C,IAAI,CAAC,IAAI,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,SAAS,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;QACvD,CAAC;QAED,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,EAAE,WAAW,CAAC,MAAM,CAAC,CAAC;QAErF,OAAO,CAAC,GAAG,CAAC,GAAG,WAAW,CAAC,MAAM,CAAC,MAAM,CAAC,aAAa,CAAC,CAAC;QACxD,OAAO,CAAC,GAAG,CAAC,GAAG,GAAG,CAAC,MAAM,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC;QACxD,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;YACvB,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;QACnE,CAAC;QAED,OAAO,CAAC,GAAG,CAAC,YAAY,UAAU,CAAC,MAAM,oBAAoB,CAAC,CAAC;IACjE,CAAC,CAAC,CAAC;AACP,CAAC"}
|
package/dist/commands/serve.d.ts
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"serve.d.ts","sourceRoot":"","sources":["../../src/commands/serve.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAIpC,wBAAgB,oBAAoB,CAAC,OAAO,EAAE,OAAO,GAAG,IAAI,
|
|
1
|
+
{"version":3,"file":"serve.d.ts","sourceRoot":"","sources":["../../src/commands/serve.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAIpC,wBAAgB,oBAAoB,CAAC,OAAO,EAAE,OAAO,GAAG,IAAI,CA+F3D"}
|
package/dist/commands/serve.js
CHANGED
|
@@ -1,21 +1,21 @@
|
|
|
1
|
-
import { createServer } from
|
|
2
|
-
import { MarkItDown } from
|
|
1
|
+
import { createServer } from "http";
|
|
2
|
+
import { MarkItDown } from "@markitdownjs/core";
|
|
3
3
|
export function registerServeCommand(program) {
|
|
4
4
|
program
|
|
5
|
-
.command(
|
|
6
|
-
.description(
|
|
7
|
-
.option(
|
|
8
|
-
.option(
|
|
5
|
+
.command("serve")
|
|
6
|
+
.description("Start an HTTP server for document conversion")
|
|
7
|
+
.option("-p, --port <port>", "Port to listen on", "3000")
|
|
8
|
+
.option("-v, --verbose", "Verbose output")
|
|
9
9
|
.action((options) => {
|
|
10
10
|
const port = parseInt(options.port, 10);
|
|
11
11
|
const parser = new MarkItDown();
|
|
12
12
|
const server = createServer(async (req, res) => {
|
|
13
|
-
if (req.method ===
|
|
14
|
-
res.writeHead(200, {
|
|
15
|
-
res.end(JSON.stringify({ status:
|
|
13
|
+
if (req.method === "GET" && req.url === "/health") {
|
|
14
|
+
res.writeHead(200, { "Content-Type": "application/json" });
|
|
15
|
+
res.end(JSON.stringify({ status: "ok", timestamp: new Date().toISOString() }));
|
|
16
16
|
return;
|
|
17
17
|
}
|
|
18
|
-
if (req.method ===
|
|
18
|
+
if (req.method === "POST" && req.url === "/convert") {
|
|
19
19
|
try {
|
|
20
20
|
const chunks = [];
|
|
21
21
|
for await (const chunk of req) {
|
|
@@ -24,24 +24,24 @@ export function registerServeCommand(program) {
|
|
|
24
24
|
const body = Buffer.concat(chunks).toString();
|
|
25
25
|
let fileData;
|
|
26
26
|
let fileName;
|
|
27
|
-
const contentType = req.headers[
|
|
28
|
-
if (contentType.includes(
|
|
27
|
+
const contentType = req.headers["content-type"] ?? "";
|
|
28
|
+
if (contentType.includes("multipart/form-data")) {
|
|
29
29
|
const boundaryMatch = contentType.match(/boundary=(.+)/);
|
|
30
30
|
if (!boundaryMatch) {
|
|
31
|
-
res.writeHead(400, {
|
|
32
|
-
res.end(JSON.stringify({ error:
|
|
31
|
+
res.writeHead(400, { "Content-Type": "application/json" });
|
|
32
|
+
res.end(JSON.stringify({ error: "Missing boundary in multipart content-type" }));
|
|
33
33
|
return;
|
|
34
34
|
}
|
|
35
35
|
const boundary = boundaryMatch[1];
|
|
36
36
|
const parts = body.split(`--${boundary}`);
|
|
37
37
|
for (const part of parts) {
|
|
38
|
-
if (part.includes(
|
|
38
|
+
if (part.includes("Content-Disposition")) {
|
|
39
39
|
const nameMatch = part.match(/name="([^"]+)"/);
|
|
40
40
|
const filenameMatch = part.match(/filename="([^"]+)"/);
|
|
41
|
-
if (nameMatch?.[1] ===
|
|
42
|
-
const headerEnd = part.indexOf(
|
|
41
|
+
if (nameMatch?.[1] === "file") {
|
|
42
|
+
const headerEnd = part.indexOf("\r\n\r\n");
|
|
43
43
|
if (headerEnd !== -1) {
|
|
44
|
-
fileData = part.substring(headerEnd + 4).replace(/\r\n--$/,
|
|
44
|
+
fileData = part.substring(headerEnd + 4).replace(/\r\n--$/, "");
|
|
45
45
|
if (filenameMatch?.[1]) {
|
|
46
46
|
fileName = filenameMatch[1];
|
|
47
47
|
}
|
|
@@ -56,30 +56,30 @@ export function registerServeCommand(program) {
|
|
|
56
56
|
fileName = json.fileName;
|
|
57
57
|
}
|
|
58
58
|
if (!fileData) {
|
|
59
|
-
res.writeHead(400, {
|
|
60
|
-
res.end(JSON.stringify({ error:
|
|
59
|
+
res.writeHead(400, { "Content-Type": "application/json" });
|
|
60
|
+
res.end(JSON.stringify({ error: "No file provided" }));
|
|
61
61
|
return;
|
|
62
62
|
}
|
|
63
63
|
const result = await parser.convert({
|
|
64
|
-
data: Buffer.from(fileData,
|
|
64
|
+
data: Buffer.from(fileData, "base64"),
|
|
65
65
|
fileName,
|
|
66
66
|
});
|
|
67
|
-
res.writeHead(200, {
|
|
67
|
+
res.writeHead(200, { "Content-Type": "application/json" });
|
|
68
68
|
res.end(JSON.stringify(result));
|
|
69
69
|
}
|
|
70
70
|
catch (error) {
|
|
71
71
|
if (options.verbose) {
|
|
72
|
-
console.error(
|
|
72
|
+
console.error("Conversion error:", error);
|
|
73
73
|
}
|
|
74
|
-
res.writeHead(500, {
|
|
74
|
+
res.writeHead(500, { "Content-Type": "application/json" });
|
|
75
75
|
res.end(JSON.stringify({
|
|
76
|
-
error: error instanceof Error ? error.message :
|
|
76
|
+
error: error instanceof Error ? error.message : "Conversion failed",
|
|
77
77
|
}));
|
|
78
78
|
}
|
|
79
79
|
return;
|
|
80
80
|
}
|
|
81
|
-
res.writeHead(404, {
|
|
82
|
-
res.end(JSON.stringify({ error:
|
|
81
|
+
res.writeHead(404, { "Content-Type": "application/json" });
|
|
82
|
+
res.end(JSON.stringify({ error: "Not found" }));
|
|
83
83
|
});
|
|
84
84
|
server.listen(port, () => {
|
|
85
85
|
console.error(`MarkItDownJS server running on http://localhost:${port}`);
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"serve.js","sourceRoot":"","sources":["../../src/commands/serve.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,YAAY,EAAmC,MAAM,MAAM,CAAC;AACrE,OAAO,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAEhD,MAAM,UAAU,oBAAoB,CAAC,OAAgB;IACnD,OAAO;SACJ,OAAO,CAAC,OAAO,CAAC;SAChB,WAAW,CAAC,8CAA8C,CAAC;SAC3D,MAAM,CAAC,mBAAmB,EAAE,mBAAmB,EAAE,MAAM,CAAC;SACxD,MAAM,CAAC,eAAe,EAAE,gBAAgB,CAAC;SACzC,MAAM,CAAC,CAAC,OAA2C,EAAE,EAAE;QACtD,MAAM,IAAI,GAAG,QAAQ,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;QACxC,MAAM,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;QAEhC,MAAM,MAAM,GAAG,YAAY,CAAC,KAAK,EAAE,GAAoB,EAAE,GAAmB,EAAE,EAAE;YAC9E,IAAI,GAAG,CAAC,MAAM,KAAK,KAAK,IAAI,GAAG,CAAC,GAAG,KAAK,SAAS,EAAE,CAAC;gBAClD,GAAG,CAAC,SAAS,CAAC,GAAG,EAAE,EAAE,cAAc,EAAE,kBAAkB,EAAE,CAAC,CAAC;gBAC3D,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,MAAM,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC,CAAC;gBAC/E,OAAO;YACT,CAAC;YAED,IAAI,GAAG,CAAC,MAAM,KAAK,MAAM,IAAI,GAAG,CAAC,GAAG,KAAK,UAAU,EAAE,CAAC;gBACpD,IAAI,CAAC;oBACH,MAAM,MAAM,GAAa,EAAE,CAAC;oBAC5B,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,GAAG,EAAE,CAAC;wBAC9B,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;oBACrB,CAAC;oBACD,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,QAAQ,EAAE,CAAC;oBAE9C,IAAI,QAA4B,CAAC;oBACjC,IAAI,QAA4B,CAAC;oBAEjC,MAAM,WAAW,GAAG,GAAG,CAAC,OAAO,CAAC,cAAc,CAAC,IAAI,EAAE,CAAC;oBACtD,IAAI,WAAW,CAAC,QAAQ,CAAC,qBAAqB,CAAC,EAAE,CAAC;wBAChD,MAAM,aAAa,GAAG,WAAW,CAAC,KAAK,CAAC,eAAe,CAAC,CAAC;wBACzD,IAAI,CAAC,aAAa,EAAE,CAAC;4BACnB,GAAG,CAAC,SAAS,CAAC,GAAG,EAAE,EAAE,cAAc,EAAE,kBAAkB,EAAE,CAAC,CAAC;4BAC3D,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,KAAK,EAAE,4CAA4C,EAAE,CAAC,CAAC,CAAC;4BACjF,OAAO;wBACT,CAAC;wBAED,MAAM,QAAQ,GAAG,aAAa,CAAC,CAAC,CAAE,CAAC;wBACnC,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,QAAQ,EAAE,CAAC,CAAC;wBAC1C,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;4BACzB,IAAI,IAAI,CAAC,QAAQ,CAAC,qBAAqB,CAAC,EAAE,CAAC;gCACzC,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,gBAAgB,CAAC,CAAC;gCAC/C,MAAM,aAAa,GAAG,IAAI,CAAC,KAAK,CAAC,oBAAoB,CAAC,CAAC;gCACvD,IAAI,SAAS,EAAE,CAAC,CAAC,CAAC,KAAK,MAAM,EAAE,CAAC;oCAC9B,MAAM,SAAS,GAAG,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC;oCAC3C,IAAI,SAAS,KAAK,CAAC,CAAC,EAAE,CAAC;wCACrB,QAAQ,GAAG,IAAI,CAAC,SAAS,CAAC,SAAS,GAAG,CAAC,CAAC,CAAC,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC,CAAC;wCAChE,IAAI,aAAa,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;4CACvB,QAAQ,GAAG,aAAa,CAAC,CAAC,CAAC,CAAC;wCAC9B,CAAC;oCACH,CAAC;gCACH,CAAC;4BACH,CAAC;wBACH,CAAC;oBACH,CAAC;yBAAM,CAAC;wBACN,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;wBAC9B,QAAQ,GAAG,IAAI,CAAC,IAAI,CAAC;wBACrB,QAAQ,GAAG,IAAI,CAAC,QAAQ,CAAC;oBAC3B,CAAC;oBAED,IAAI,CAAC,QAAQ,EAAE,CAAC;wBACd,GAAG,CAAC,SAAS,CAAC,GAAG,EAAE,EAAE,cAAc,EAAE,kBAAkB,EAAE,CAAC,CAAC;wBAC3D,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,KAAK,EAAE,kBAAkB,EAAE,CAAC,CAAC,CAAC;wBACvD,OAAO;oBACT,CAAC;oBAED,MAAM,MAAM,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC;wBAClC,IAAI,EAAE,MAAM,CAAC,IAAI,CAAC,QAAQ,EAAE,QAAQ,CAAC;wBACrC,QAAQ;qBACT,CAAC,CAAC;oBAEH,GAAG,CAAC,SAAS,CAAC,GAAG,EAAE,EAAE,cAAc,EAAE,kBAAkB,EAAE,CAAC,CAAC;oBAC3D,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC;gBAClC,CAAC;gBAAC,OAAO,KAAK,EAAE,CAAC;oBACf,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;wBACpB,OAAO,CAAC,KAAK,CAAC,mBAAmB,EAAE,KAAK,CAAC,CAAC;oBAC5C,CAAC;oBACD,GAAG,CAAC,SAAS,CAAC,GAAG,EAAE,EAAE,cAAc,EAAE,kBAAkB,EAAE,CAAC,CAAC;oBAC3D,GAAG,CAAC,GAAG,
|
|
1
|
+
{"version":3,"file":"serve.js","sourceRoot":"","sources":["../../src/commands/serve.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,YAAY,EAAmC,MAAM,MAAM,CAAC;AACrE,OAAO,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAEhD,MAAM,UAAU,oBAAoB,CAAC,OAAgB;IACnD,OAAO;SACJ,OAAO,CAAC,OAAO,CAAC;SAChB,WAAW,CAAC,8CAA8C,CAAC;SAC3D,MAAM,CAAC,mBAAmB,EAAE,mBAAmB,EAAE,MAAM,CAAC;SACxD,MAAM,CAAC,eAAe,EAAE,gBAAgB,CAAC;SACzC,MAAM,CAAC,CAAC,OAA2C,EAAE,EAAE;QACtD,MAAM,IAAI,GAAG,QAAQ,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;QACxC,MAAM,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;QAEhC,MAAM,MAAM,GAAG,YAAY,CAAC,KAAK,EAAE,GAAoB,EAAE,GAAmB,EAAE,EAAE;YAC9E,IAAI,GAAG,CAAC,MAAM,KAAK,KAAK,IAAI,GAAG,CAAC,GAAG,KAAK,SAAS,EAAE,CAAC;gBAClD,GAAG,CAAC,SAAS,CAAC,GAAG,EAAE,EAAE,cAAc,EAAE,kBAAkB,EAAE,CAAC,CAAC;gBAC3D,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,MAAM,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC,CAAC;gBAC/E,OAAO;YACT,CAAC;YAED,IAAI,GAAG,CAAC,MAAM,KAAK,MAAM,IAAI,GAAG,CAAC,GAAG,KAAK,UAAU,EAAE,CAAC;gBACpD,IAAI,CAAC;oBACH,MAAM,MAAM,GAAa,EAAE,CAAC;oBAC5B,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,GAAG,EAAE,CAAC;wBAC9B,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;oBACrB,CAAC;oBACD,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,QAAQ,EAAE,CAAC;oBAE9C,IAAI,QAA4B,CAAC;oBACjC,IAAI,QAA4B,CAAC;oBAEjC,MAAM,WAAW,GAAG,GAAG,CAAC,OAAO,CAAC,cAAc,CAAC,IAAI,EAAE,CAAC;oBACtD,IAAI,WAAW,CAAC,QAAQ,CAAC,qBAAqB,CAAC,EAAE,CAAC;wBAChD,MAAM,aAAa,GAAG,WAAW,CAAC,KAAK,CAAC,eAAe,CAAC,CAAC;wBACzD,IAAI,CAAC,aAAa,EAAE,CAAC;4BACnB,GAAG,CAAC,SAAS,CAAC,GAAG,EAAE,EAAE,cAAc,EAAE,kBAAkB,EAAE,CAAC,CAAC;4BAC3D,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,KAAK,EAAE,4CAA4C,EAAE,CAAC,CAAC,CAAC;4BACjF,OAAO;wBACT,CAAC;wBAED,MAAM,QAAQ,GAAG,aAAa,CAAC,CAAC,CAAE,CAAC;wBACnC,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,QAAQ,EAAE,CAAC,CAAC;wBAC1C,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;4BACzB,IAAI,IAAI,CAAC,QAAQ,CAAC,qBAAqB,CAAC,EAAE,CAAC;gCACzC,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,gBAAgB,CAAC,CAAC;gCAC/C,MAAM,aAAa,GAAG,IAAI,CAAC,KAAK,CAAC,oBAAoB,CAAC,CAAC;gCACvD,IAAI,SAAS,EAAE,CAAC,CAAC,CAAC,KAAK,MAAM,EAAE,CAAC;oCAC9B,MAAM,SAAS,GAAG,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC;oCAC3C,IAAI,SAAS,KAAK,CAAC,CAAC,EAAE,CAAC;wCACrB,QAAQ,GAAG,IAAI,CAAC,SAAS,CAAC,SAAS,GAAG,CAAC,CAAC,CAAC,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC,CAAC;wCAChE,IAAI,aAAa,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;4CACvB,QAAQ,GAAG,aAAa,CAAC,CAAC,CAAC,CAAC;wCAC9B,CAAC;oCACH,CAAC;gCACH,CAAC;4BACH,CAAC;wBACH,CAAC;oBACH,CAAC;yBAAM,CAAC;wBACN,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;wBAC9B,QAAQ,GAAG,IAAI,CAAC,IAAI,CAAC;wBACrB,QAAQ,GAAG,IAAI,CAAC,QAAQ,CAAC;oBAC3B,CAAC;oBAED,IAAI,CAAC,QAAQ,EAAE,CAAC;wBACd,GAAG,CAAC,SAAS,CAAC,GAAG,EAAE,EAAE,cAAc,EAAE,kBAAkB,EAAE,CAAC,CAAC;wBAC3D,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,KAAK,EAAE,kBAAkB,EAAE,CAAC,CAAC,CAAC;wBACvD,OAAO;oBACT,CAAC;oBAED,MAAM,MAAM,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC;wBAClC,IAAI,EAAE,MAAM,CAAC,IAAI,CAAC,QAAQ,EAAE,QAAQ,CAAC;wBACrC,QAAQ;qBACT,CAAC,CAAC;oBAEH,GAAG,CAAC,SAAS,CAAC,GAAG,EAAE,EAAE,cAAc,EAAE,kBAAkB,EAAE,CAAC,CAAC;oBAC3D,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC;gBAClC,CAAC;gBAAC,OAAO,KAAK,EAAE,CAAC;oBACf,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;wBACpB,OAAO,CAAC,KAAK,CAAC,mBAAmB,EAAE,KAAK,CAAC,CAAC;oBAC5C,CAAC;oBACD,GAAG,CAAC,SAAS,CAAC,GAAG,EAAE,EAAE,cAAc,EAAE,kBAAkB,EAAE,CAAC,CAAC;oBAC3D,GAAG,CAAC,GAAG,CACL,IAAI,CAAC,SAAS,CAAC;wBACb,KAAK,EAAE,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,mBAAmB;qBACpE,CAAC,CACH,CAAC;gBACJ,CAAC;gBACD,OAAO;YACT,CAAC;YAED,GAAG,CAAC,SAAS,CAAC,GAAG,EAAE,EAAE,cAAc,EAAE,kBAAkB,EAAE,CAAC,CAAC;YAC3D,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,KAAK,EAAE,WAAW,EAAE,CAAC,CAAC,CAAC;QAClD,CAAC,CAAC,CAAC;QAEH,MAAM,CAAC,MAAM,CAAC,IAAI,EAAE,GAAG,EAAE;YACvB,OAAO,CAAC,KAAK,CAAC,mDAAmD,IAAI,EAAE,CAAC,CAAC;QAC3E,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;AACP,CAAC"}
|
package/dist/commands/watch.d.ts
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"watch.d.ts","sourceRoot":"","sources":["../../src/commands/watch.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAKpC,wBAAgB,oBAAoB,CAAC,OAAO,EAAE,OAAO,GAAG,IAAI,
|
|
1
|
+
{"version":3,"file":"watch.d.ts","sourceRoot":"","sources":["../../src/commands/watch.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAKpC,wBAAgB,oBAAoB,CAAC,OAAO,EAAE,OAAO,GAAG,IAAI,CAqF3D"}
|
package/dist/commands/watch.js
CHANGED
|
@@ -1,13 +1,13 @@
|
|
|
1
|
-
import { watch, readFile, writeFile, stat, readdir } from
|
|
2
|
-
import { join, basename, extname, resolve } from
|
|
3
|
-
import { MarkItDown } from
|
|
1
|
+
import { watch, readFile, writeFile, stat, readdir } from "fs/promises";
|
|
2
|
+
import { join, basename, extname, resolve } from "path";
|
|
3
|
+
import { MarkItDown } from "@markitdownjs/core";
|
|
4
4
|
export function registerWatchCommand(program) {
|
|
5
5
|
program
|
|
6
|
-
.command(
|
|
7
|
-
.description(
|
|
8
|
-
.argument(
|
|
9
|
-
.option(
|
|
10
|
-
.option(
|
|
6
|
+
.command("watch")
|
|
7
|
+
.description("Watch a directory and convert new/modified files to Markdown")
|
|
8
|
+
.argument("[dir]", "Directory to watch", ".")
|
|
9
|
+
.option("-r, --recursive", "Watch subdirectories recursively")
|
|
10
|
+
.option("-v, --verbose", "Verbose output")
|
|
11
11
|
.action(async (dir, options) => {
|
|
12
12
|
const watchDir = resolve(dir);
|
|
13
13
|
const parser = new MarkItDown();
|
|
@@ -20,8 +20,8 @@ export function registerWatchCommand(program) {
|
|
|
20
20
|
data: new Uint8Array(content),
|
|
21
21
|
fileName: filePath,
|
|
22
22
|
});
|
|
23
|
-
const mdPath = join(watchDir, basename(filePath, extname(filePath)) +
|
|
24
|
-
await writeFile(mdPath, result.markdown,
|
|
23
|
+
const mdPath = join(watchDir, basename(filePath, extname(filePath)) + ".md");
|
|
24
|
+
await writeFile(mdPath, result.markdown, "utf-8");
|
|
25
25
|
if (options.verbose) {
|
|
26
26
|
console.error(`Converted: ${filePath} -> ${mdPath}`);
|
|
27
27
|
}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"watch.js","sourceRoot":"","sources":["../../src/commands/watch.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,KAAK,EAAE,QAAQ,EAAE,SAAS,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,aAAa,CAAC;AACxE,OAAO,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM,MAAM,CAAC;AACxD,OAAO,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAEhD,MAAM,UAAU,oBAAoB,CAAC,OAAgB;IACnD,OAAO;SACJ,OAAO,CAAC,OAAO,CAAC;SAChB,WAAW,CAAC,8DAA8D,CAAC;SAC3E,QAAQ,CAAC,OAAO,EAAE,oBAAoB,EAAE,GAAG,CAAC;SAC5C,MAAM,CAAC,iBAAiB,EAAE,kCAAkC,CAAC;SAC7D,MAAM,CAAC,eAAe,EAAE,gBAAgB,CAAC;SACzC,MAAM,CAAC,KAAK,EAAE,GAAW,EAAE,OAAiD,EAAE,EAAE;QAC/E,MAAM,QAAQ,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC;QAC9B,MAAM,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;QAChC,MAAM,cAAc,GAAG,IAAI,GAAG,EAAU,CAAC;QAEzC,OAAO,CAAC,KAAK,CAAC,aAAa,QAAQ,EAAE,CAAC,CAAC;QAEvC,KAAK,UAAU,WAAW,CAAC,QAAgB;YACzC,IAAI,CAAC;gBACH,MAAM,OAAO,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAC;gBACzC,MAAM,MAAM,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC;oBAClC,IAAI,EAAE,IAAI,UAAU,CAAC,OAAO,CAAC;oBAC7B,QAAQ,EAAE,QAAQ;iBACnB,CAAC,CAAC;gBAEH,MAAM,MAAM,GAAG,IAAI,
|
|
1
|
+
{"version":3,"file":"watch.js","sourceRoot":"","sources":["../../src/commands/watch.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,KAAK,EAAE,QAAQ,EAAE,SAAS,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,aAAa,CAAC;AACxE,OAAO,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM,MAAM,CAAC;AACxD,OAAO,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAEhD,MAAM,UAAU,oBAAoB,CAAC,OAAgB;IACnD,OAAO;SACJ,OAAO,CAAC,OAAO,CAAC;SAChB,WAAW,CAAC,8DAA8D,CAAC;SAC3E,QAAQ,CAAC,OAAO,EAAE,oBAAoB,EAAE,GAAG,CAAC;SAC5C,MAAM,CAAC,iBAAiB,EAAE,kCAAkC,CAAC;SAC7D,MAAM,CAAC,eAAe,EAAE,gBAAgB,CAAC;SACzC,MAAM,CAAC,KAAK,EAAE,GAAW,EAAE,OAAiD,EAAE,EAAE;QAC/E,MAAM,QAAQ,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC;QAC9B,MAAM,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;QAChC,MAAM,cAAc,GAAG,IAAI,GAAG,EAAU,CAAC;QAEzC,OAAO,CAAC,KAAK,CAAC,aAAa,QAAQ,EAAE,CAAC,CAAC;QAEvC,KAAK,UAAU,WAAW,CAAC,QAAgB;YACzC,IAAI,CAAC;gBACH,MAAM,OAAO,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAC;gBACzC,MAAM,MAAM,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC;oBAClC,IAAI,EAAE,IAAI,UAAU,CAAC,OAAO,CAAC;oBAC7B,QAAQ,EAAE,QAAQ;iBACnB,CAAC,CAAC;gBAEH,MAAM,MAAM,GAAG,IAAI,CAAC,QAAQ,EAAE,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,QAAQ,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC;gBAC7E,MAAM,SAAS,CAAC,MAAM,EAAE,MAAM,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;gBAElD,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;oBACpB,OAAO,CAAC,KAAK,CAAC,cAAc,QAAQ,OAAO,MAAM,EAAE,CAAC,CAAC;gBACvD,CAAC;YACH,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,OAAO,CAAC,KAAK,CACX,oBAAoB,QAAQ,KAAK,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAC1F,CAAC;YACJ,CAAC;QACH,CAAC;QAED,KAAK,UAAU,aAAa,CAAC,OAAe;YAC1C,IAAI,CAAC;gBACH,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,OAAO,EAAE,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC,CAAC;gBAChE,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;oBAC5B,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC;oBAC3C,IAAI,KAAK,CAAC,WAAW,EAAE,IAAI,OAAO,CAAC,SAAS,EAAE,CAAC;wBAC7C,MAAM,aAAa,CAAC,QAAQ,CAAC,CAAC;oBAChC,CAAC;yBAAM,IAAI,KAAK,CAAC,MAAM,EAAE,EAAE,CAAC;wBAC1B,cAAc,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;wBAC7B,MAAM,WAAW,CAAC,QAAQ,CAAC,CAAC;oBAC9B,CAAC;gBACH,CAAC;YACH,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,OAAO,CAAC,KAAK,CACX,kBAAkB,OAAO,KAAK,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CACvF,CAAC;YACJ,CAAC;QACH,CAAC;QAED,MAAM,aAAa,CAAC,QAAQ,CAAC,CAAC;QAE9B,IAAI,CAAC;YACH,MAAM,OAAO,GAAG,KAAK,CAAC,QAAQ,EAAE,EAAE,SAAS,EAAE,OAAO,CAAC,SAAS,EAAE,CAAC,CAAC;YAElE,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;gBAClC,IAAI,CAAC,KAAK,CAAC,QAAQ;oBAAE,SAAS;gBAE9B,MAAM,QAAQ,GAAG,IAAI,CAAC,QAAQ,EAAE,KAAK,CAAC,QAAQ,CAAC,CAAC;gBAEhD,IAAI,CAAC;oBACH,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,CAAC;oBACtC,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE;wBAAE,SAAS;gBACnC,CAAC;gBAAC,MAAM,CAAC;oBACP,SAAS;gBACX,CAAC;gBAED,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC;oBAClC,cAAc,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;oBAC7B,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;wBACpB,OAAO,CAAC,KAAK,CAAC,sBAAsB,QAAQ,EAAE,CAAC,CAAC;oBAClD,CAAC;gBACH,CAAC;gBAED,MAAM,WAAW,CAAC,QAAQ,CAAC,CAAC;YAC9B,CAAC;QACH,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,OAAO,CAAC,KAAK,CAAC,gBAAgB,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;YACxF,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAClB,CAAC;IACH,CAAC,CAAC,CAAC;AACP,CAAC"}
|
package/dist/index.js
CHANGED
|
@@ -1,15 +1,15 @@
|
|
|
1
1
|
#!/usr/bin/env node
|
|
2
|
-
import { Command } from
|
|
3
|
-
import { registerConvertCommand } from
|
|
4
|
-
import { registerWatchCommand } from
|
|
5
|
-
import { registerBatchCommand } from
|
|
6
|
-
import { registerServeCommand } from
|
|
7
|
-
import { registerFormatsCommand } from
|
|
2
|
+
import { Command } from "commander";
|
|
3
|
+
import { registerConvertCommand } from "./commands/convert.js";
|
|
4
|
+
import { registerWatchCommand } from "./commands/watch.js";
|
|
5
|
+
import { registerBatchCommand } from "./commands/batch.js";
|
|
6
|
+
import { registerServeCommand } from "./commands/serve.js";
|
|
7
|
+
import { registerFormatsCommand } from "./commands/formats.js";
|
|
8
8
|
const program = new Command();
|
|
9
9
|
program
|
|
10
|
-
.name(
|
|
11
|
-
.description(
|
|
12
|
-
.version(
|
|
10
|
+
.name("markitdownjs")
|
|
11
|
+
.description("Universal document-to-Markdown conversion tool")
|
|
12
|
+
.version("0.1.0");
|
|
13
13
|
registerConvertCommand(program);
|
|
14
14
|
registerWatchCommand(program);
|
|
15
15
|
registerBatchCommand(program);
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@markitdownjs/cli",
|
|
3
|
-
"version": "0.1.
|
|
3
|
+
"version": "0.1.3",
|
|
4
4
|
"description": "CLI tool for MarkItDownJS document conversion",
|
|
5
5
|
"license": "MIT",
|
|
6
6
|
"type": "module",
|
|
@@ -18,12 +18,12 @@
|
|
|
18
18
|
],
|
|
19
19
|
"dependencies": {
|
|
20
20
|
"commander": "^12.0.0",
|
|
21
|
-
"@markitdownjs/core": "0.
|
|
22
|
-
"@markitdownjs/shared": "0.
|
|
21
|
+
"@markitdownjs/core": "0.3.0",
|
|
22
|
+
"@markitdownjs/shared": "0.3.0"
|
|
23
23
|
},
|
|
24
24
|
"devDependencies": {
|
|
25
25
|
"typescript": "^5.5.0",
|
|
26
|
-
"@types/node": "^20.
|
|
26
|
+
"@types/node": "^20.19.43"
|
|
27
27
|
},
|
|
28
28
|
"scripts": {
|
|
29
29
|
"build": "tsc --project tsconfig.json",
|