20261008

uniCloud-tcb/cloudfunctions/router/node_modules/pdf-parse/lib/pdf-parse.js

@@ -0,0 +1,102 @@
+var PDFJS = null;
+
+function render_page(pageData) {
+    //check documents https://mozilla.github.io/pdf.js/
+    //ret.text = ret.text ? ret.text : "";
+
+    let render_options = {
+        //replaces all occurrences of whitespace with standard spaces (0x20). The default value is `false`.
+        normalizeWhitespace: false,
+        //do not attempt to combine same line TextItem's. The default value is `false`.
+        disableCombineTextItems: false
+    }
+
+    return pageData.getTextContent(render_options)
+        .then(function(textContent) {
+            let lastY, text = '';
+            //https://github.com/mozilla/pdf.js/issues/8963
+            //https://github.com/mozilla/pdf.js/issues/2140
+            //https://gist.github.com/hubgit/600ec0c224481e910d2a0f883a7b98e3
+            //https://gist.github.com/hubgit/600ec0c224481e910d2a0f883a7b98e3
+            for (let item of textContent.items) {
+                if (lastY == item.transform[5] || !lastY){
+                    text += item.str;
+                }  
+                else{
+                    text += '\n' + item.str;
+                }    
+                lastY = item.transform[5];
+            }            
+            //let strings = textContent.items.map(item => item.str);
+            //let text = strings.join("\n");
+            //text = text.replace(/[ ]+/ig," ");
+            //ret.text = `${ret.text} ${text} \n\n`;
+            return text;
+        });
+}
+
+const DEFAULT_OPTIONS = {
+    pagerender: render_page,
+    max: 0,
+    //check https://mozilla.github.io/pdf.js/getting_started/
+    version: 'v1.10.100'
+}
+
+async function PDF(dataBuffer, options) {
+    var isDebugMode = false;
+
+    let ret = {
+        numpages: 0,
+        numrender: 0,
+        info: null,
+        metadata: null,
+        text: "",
+        version: null
+    };
+
+    if (typeof options == 'undefined') options = DEFAULT_OPTIONS;
+    if (typeof options.pagerender != 'function') options.pagerender = DEFAULT_OPTIONS.pagerender;
+    if (typeof options.max != 'number') options.max = DEFAULT_OPTIONS.max;
+    if (typeof options.version != 'string') options.version = DEFAULT_OPTIONS.version;
+    if (options.version == 'default') options.version = DEFAULT_OPTIONS.version;
+
+    PDFJS = PDFJS ? PDFJS : require(`./pdf.js/${options.version}/build/pdf.js`);
+
+    ret.version = PDFJS.version;
+
+    // Disable workers to avoid yet another cross-origin issue (workers need
+    // the URL of the script to be loaded, and dynamically loading a cross-origin
+    // script does not work).
+    PDFJS.disableWorker = true;
+    let doc = await PDFJS.getDocument(dataBuffer);
+    ret.numpages = doc.numPages;
+
+    let metaData = await doc.getMetadata().catch(function(err) {
+        return null;
+    });
+
+    ret.info = metaData ? metaData.info : null;
+    ret.metadata = metaData ? metaData.metadata : null;
+
+    let counter = options.max <= 0 ? doc.numPages : options.max;
+    counter = counter > doc.numPages ? doc.numPages : counter;
+
+    ret.text = "";
+
+    for (var i = 1; i <= counter; i++) {
+        let pageText = await doc.getPage(i).then(pageData => options.pagerender(pageData)).catch((err)=>{
+            // todo log err using debug
+            debugger;
+            return "";
+        });
+
+        ret.text = `${ret.text}\n\n${pageText}`;
+    }
+
+    ret.numrender = counter;
+    doc.destroy();
+
+    return ret;
+}
+
+module.exports = PDF;

uniCloud-tcb/cloudfunctions/router/node_modules/pdf-parse/lib/pdf.js/README.md

@@ -0,0 +1,2 @@
+# check https://mozilla.github.io/pdf.js/getting_started/#download 
+# just download then delete web folder and license file

uniCloud-tcb/cloudfunctions/router/node_modules/pdf-parse/lib/pdf.js/v1.10.88/web/cmaps/CNS2-V.bcmap

@@ -0,0 +1,3 @@
+àRCopyright 1990-2009 Adobe Systems Incorporated.
+All rights reserved.
+See ./LICENSEáCNS2-H

uniCloud-tcb/cloudfunctions/router/node_modules/pdf-parse/lib/pdf.js/v1.10.88/web/cmaps/ETenms-B5-H.bcmap

@@ -0,0 +1,3 @@
+àRCopyright 1990-2009 Adobe Systems Incorporated.
+All rights reserved.
+See ./LICENSEá	ETen-B5-H`
 ^

uniCloud-tcb/cloudfunctions/router/node_modules/pdf-parse/lib/pdf.js/v1.10.88/web/cmaps/GB-H.bcmap

@@ -0,0 +1,4 @@
+àRCopyright 1990-2009 Adobe Systems Incorporated.
+All rights reserved.
+See ./LICENSE

!!�º]aX!!]`�21�>	�p�z�$]‚�"R‚d�-Uƒ7�*„
+4„%�+ „Z „{�/…%…<�9K…b�1]†.�"‡‰`]‡,�"]ˆ