@veryfront/ext-document-kreuzberg 0.1.993 → 0.1.994
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/esm/extensions/ext-document-kreuzberg/src/extraction-config.d.ts +1 -1
- package/esm/extensions/ext-document-kreuzberg/src/extraction-config.d.ts.map +1 -1
- package/esm/extensions/ext-document-kreuzberg/src/extraction-config.js +7 -3
- package/esm/extensions/ext-document-kreuzberg/src/native-progress-extraction-worker.js +10 -2
- package/package.json +2 -2
|
@@ -1,2 +1,2 @@
|
|
|
1
|
-
export declare function extractionConfigForMimeType(mimeType: string): Record<string, unknown
|
|
1
|
+
export declare function extractionConfigForMimeType(mimeType: string): Record<string, unknown>;
|
|
2
2
|
//# sourceMappingURL=extraction-config.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"extraction-config.d.ts","sourceRoot":"","sources":["../../../../src/extensions/ext-document-kreuzberg/src/extraction-config.ts"],"names":[],"mappings":"
|
|
1
|
+
{"version":3,"file":"extraction-config.d.ts","sourceRoot":"","sources":["../../../../src/extensions/ext-document-kreuzberg/src/extraction-config.ts"],"names":[],"mappings":"AAmBA,wBAAgB,2BAA2B,CAAC,QAAQ,EAAE,MAAM,GAAG,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAIrF"}
|
|
@@ -1,4 +1,8 @@
|
|
|
1
|
-
const
|
|
1
|
+
const MARKDOWN_EXTRACTION_CONFIG = {
|
|
2
|
+
outputFormat: "markdown",
|
|
3
|
+
};
|
|
4
|
+
const PDF_MARKDOWN_EXTRACTION_CONFIG = {
|
|
5
|
+
...MARKDOWN_EXTRACTION_CONFIG,
|
|
2
6
|
images: { extractImages: false },
|
|
3
7
|
pdfOptions: {
|
|
4
8
|
extractImages: false,
|
|
@@ -12,6 +16,6 @@ function normalizeMimeType(mimeType) {
|
|
|
12
16
|
}
|
|
13
17
|
export function extractionConfigForMimeType(mimeType) {
|
|
14
18
|
return normalizeMimeType(mimeType) === "application/pdf"
|
|
15
|
-
?
|
|
16
|
-
:
|
|
19
|
+
? PDF_MARKDOWN_EXTRACTION_CONFIG
|
|
20
|
+
: MARKDOWN_EXTRACTION_CONFIG;
|
|
17
21
|
}
|
|
@@ -121,7 +121,11 @@ async function extractPptxBySlide(buffer, mimeType) {
|
|
|
121
121
|
const slidePaths = await pptxSlidePaths(zip);
|
|
122
122
|
if (!slidePaths.length) {
|
|
123
123
|
const { extractBytes } = await loadKreuzbergNative();
|
|
124
|
-
const result = await extractBytes(
|
|
124
|
+
const result = await extractBytes(
|
|
125
|
+
new Uint8Array(buffer),
|
|
126
|
+
mimeType,
|
|
127
|
+
extractionConfigForMimeType(mimeType)
|
|
128
|
+
);
|
|
125
129
|
postProgress({ unit: "file", current: 1, total: 1, characters: result.content.length });
|
|
126
130
|
return result.content;
|
|
127
131
|
}
|
|
@@ -142,7 +146,11 @@ async function extractPptxBySlide(buffer, mimeType) {
|
|
|
142
146
|
}
|
|
143
147
|
async function extractWholeFile(buffer, mimeType) {
|
|
144
148
|
const { extractBytes } = await loadKreuzbergNative();
|
|
145
|
-
const result = await extractBytes(
|
|
149
|
+
const result = await extractBytes(
|
|
150
|
+
new Uint8Array(buffer),
|
|
151
|
+
mimeType,
|
|
152
|
+
extractionConfigForMimeType(mimeType)
|
|
153
|
+
);
|
|
146
154
|
postProgress({ unit: "file", current: 1, total: 1, characters: result.content.length });
|
|
147
155
|
return result.content;
|
|
148
156
|
}
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@veryfront/ext-document-kreuzberg",
|
|
3
|
-
"version": "0.1.
|
|
3
|
+
"version": "0.1.994",
|
|
4
4
|
"description": "Veryfront first-party extension package for ext-document-kreuzberg",
|
|
5
5
|
"keywords": [
|
|
6
6
|
"veryfront",
|
|
@@ -55,7 +55,7 @@
|
|
|
55
55
|
"pdf-lib": "1.17.1"
|
|
56
56
|
},
|
|
57
57
|
"peerDependencies": {
|
|
58
|
-
"veryfront": "^0.1.
|
|
58
|
+
"veryfront": "^0.1.994"
|
|
59
59
|
},
|
|
60
60
|
"type": "module",
|
|
61
61
|
"types": "./esm/extensions/ext-document-kreuzberg/src/index.d.ts",
|