pdf2json - npm Package Compare versions

Comparing version 0.5.6 to 0.5.7

lib/p2jcmd.js

@@ -27,2 +27,4 @@ 'use strict';
 
+    var _iChars = "!@#$%^&*()+=[]\\\';,/{}|\":<>?~`.-_  ";
+
     var _continue = function(callback, err) {
@@ -36,38 +38,36 @@ if (err)
     var _writeOneJSON = function(data, callback) {
-        var self = this;
         var pJSON = JSON.stringify({"formImage":data});
 
-        fs.writeFile(self.outputPath, pJSON, function(err) {
+        fs.writeFile(this.outputPath, pJSON, function(err) {
             if(err) {
-                nodeUtil.p2jwarn("\n" + self.inputFile + " => " + self.outputFile + " Exception: " + err);
-                self.curProcessor.failedCount++;
+                nodeUtil.p2jwarn(this.inputFile + " => " + this.outputFile + " Exception: " + err);
+                this.curProcessor.failedCount++;
             } else {
-                nodeUtil.p2jinfo("\n" + self.inputFile + " => " + self.outputFile + " [" + self.outputDir + "] OK");
-                self.curProcessor.successCount++;
+                nodeUtil.p2jinfo(this.inputFile + " => " + this.outputFile + " [" + this.outputDir + "] OK");
+                this.curProcessor.successCount++;
             }
-            _continue.call(self, callback, err);
-        });
+            _continue.call(this, callback, err);
+        }.bind(this));
     };
 
     var _parseOnePDF = function(callback) {
-        var self = this;
-        self.pdfParser = new PFParser();
+        this.pdfParser = new PFParser();
 
-        self.pdfParser.on("pdfParser_dataReady", function (evtData) {
+        this.pdfParser.on("pdfParser_dataReady", function (evtData) {
             if ((!!evtData) && (!!evtData.data)) {
-                _writeOneJSON.call(self, evtData.data, callback);
+                _writeOneJSON.call(this, evtData.data, callback);
             }
             else {
-                self.curProcessor.failedCount++;
-                _continue.call(self, callback, "Exception: empty parsing result - " + self.inputPath);
+                this.curProcessor.failedCount++;
+                _continue.call(this, callback, "Exception: empty parsing result - " + this.inputPath);
             }
-        });
+        }.bind(this));
 
-        self.pdfParser.on("pdfParser_dataError", function (evtData) {
-            self.curProcessor.failedCount++;
+        this.pdfParser.on("pdfParser_dataError", function (evtData) {
+            this.curProcessor.failedCount++;
             var errMsg = "Exception: " + evtData.data;
-            _continue.call(self, callback, errMsg);
-        });
+            _continue.call(this, callback, errMsg);
+        }.bind(this));
 
-        self.pdfParser.loadPDF(self.inputPath, (_.has(argv, 's') ? 0 : 5));
+        this.pdfParser.loadPDF(this.inputPath, (_.has(argv, 's') ? 0 : 5));
     };
@@ -118,3 +118,3 @@
                 else {
-                    nodeUtil.p2jinfo("\nTranscoding " + this.inputFile + " to - " + this.outputPath);
+                    nodeUtil.p2jinfo("Transcoding " + this.inputFile + " to - " + this.outputPath);
                     fs.closeSync(fod);
@@ -147,3 +147,10 @@ fs.unlinkSync(this.outputPath);
         else {
-            _parseOnePDF.call(this, callback);
+            var inputFile = path.basename(this.inputFile.toLowerCase(), '.pdf');
+            if (!inputFile || inputFile.length < 1 || _iChars.indexOf(inputFile.substr(0,1)) >= 0) {
+                nodeUtil.p2jinfo("Skipped PDF " + this.inputFile + " - " + "invalid filename.");
+                _continue.call(this, callback);
+            }
+            else {
+                _parseOnePDF.call(this, callback);
+            }
         }
@@ -221,8 +228,7 @@ };
 
-        var self = this;
         process.nextTick( function() {
             console.timeEnd(_PRO_TIMER);
-            var exitCode = (self.inputCount === self.successCount) ? 0 : 1;
+            var exitCode = (this.inputCount === this.successCount) ? 0 : 1;
             process.exit(exitCode);
-        });
+        }.bind(this));
     };
@@ -240,30 +246,27 @@
     cls.prototype.processFiles = function(inputDir, files) {
-        var self = this;
         var fId = 0;
-
-        self.p2j = new PDF2JSONUtil(inputDir, files[fId], self);
-        self.p2j.processFile( function processPDFFile(err) {
+        this.p2j = new PDF2JSONUtil(inputDir, files[fId], this);
+        this.p2j.processFile( function processPDFFile(err) {
             if (err) {
-                self.complete(err);
+                this.complete(err);
             }
             else {
                 fId++;
-                if (fId >= self.inputCount) {
-                    self.complete(null);
+                if (fId >= this.inputCount) {
+                    this.complete(null);
                 }
                 else {
-                    if (self.p2j) {
-                        self.p2j.destroy();
-                        self.p2j = null;
+                    if (this.p2j) {
+                        this.p2j.destroy();
+                        this.p2j = null;
                     }
 
-                    self.p2j = new PDF2JSONUtil(inputDir, files[fId], self);
-                    self.p2j.processFile(processPDFFile);
+                    this.p2j = new PDF2JSONUtil(inputDir, files[fId], this);
+                    this.p2j.processFile(processPDFFile.bind(this));
                 }
             }
-        });
+        }.bind(this));
     };
 
     cls.prototype.processOneDirectory = function () {
-        var self = this;
         var inputDir = path.normalize(argv.f);
@@ -276,11 +279,11 @@
 
-            self.inputCount = pdfFiles.length;
-            if (self.inputCount > 0) {
-                self.processFiles(inputDir, pdfFiles);
+            this.inputCount = pdfFiles.length;
+            if (this.inputCount > 0) {
+                this.processFiles(inputDir, pdfFiles);
             }
             else {
                 console.log("No PDF files found. [" + inputDir + "].");
-                self.complete(null);
+                this.complete(null);
             }
-        });
+        }.bind(this));
     };
@@ -287,0 +290,0 @@

package.json

 {
   "name": "pdf2json",
-  "_id": "pdf2json@0.5.6",
-  "version": "0.5.6",
+  "_id": "pdf2json@0.5.7",
+  "version": "0.5.7",
   "description": "A PDF file parser that converts PDF binaries to text based JSON, powered by porting a fork of PDF.JS to Node.js",
@@ -6,0 +6,0 @@ "keywords": [

readme.md

@@ -632,2 +632,3 @@ Introduction
 
+v0.5.7 added the capability to skip input PDF files if filename begins with any one of "!@#$%^&*()+=[]\\\';,/{}|\":<>?~`.-_  ", usually these files are created by PDF authoring tools as backup files.
 
@@ -634,0 +635,0 @@ Run in a RESTful Web Service

New alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Fixed alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Improved metrics

Total package byte prevSize

3264727

increased by0.02%

Lines of code

35895

increased by0.01%

Number of lines in readme file

644

increased by0.16%

No dependency changes