fixes #106, adding epub extraction

README.md

@@ -11,6 +11,7 @@ A text extraction node module.
 * HTML, HTM
 * ATOM, RSS
 * Markdown
+* EPUB
 * XML, XSL
 * PDF
 * DOC, DOCX
@@ -170,11 +171,12 @@ textract.fromUrl(url, config, function( error, text ) {})
 
 ## Release Notes
 
-### 2.3.1 (pending)
+### 2.4.0 (pending)
 * [#164](https://github.com/dbashford/textract/issues/164). Fixed issue with extra text nodes in odt/ott extraction.
 * [#156](https://github.com/dbashford/textract/issues/156). Introduced `preserveOnlyMultipleLineBreaks` feature.
 * [#149](https://github.com/dbashford/textract/issues/149). RTF extraction error error fixed by [#166](https://github.com/dbashford/textract/pull/166).
 * [#145](https://github.com/dbashford/textract/issues/145). Handling Japanese full-width characters.
+* [#106](https://github.com/dbashford/textract/issues/106). Now extracting `.epub`
 
 156
 

lib/extractors/epub.js

@@ -0,0 +1,48 @@
+var EPub = require( 'epub2/node' )
+  , htmlExtract = require( './html' )
+  ;
+
+function extractText( filePath, options, cb ) {
+  var epub = new EPub( filePath )
+    , allText = ''
+    , hasError = false
+    , chapterCount = 0
+    ;
+
+  epub.on( 'end', function() {
+    // Iterate over each chapter...
+    epub.flow.forEach( function( chapter ) {
+      // if already error, don't do anything
+      if ( !hasError ) {
+        // Get the chapter text
+        epub.getChapterRaw( chapter.id, function( rawChaperError, text ) {
+          if ( rawChaperError ) {
+            hasError = true;
+            cb( rawChaperError, null );
+          } else {
+            // Extract the raw text from the chapter text (it's html)
+            htmlExtract.extractFromText( text, options, function( htmlExtractError, outText ) {
+              if ( htmlExtractError ) {
+                hasError = true;
+                cb( htmlExtractError, null );
+              } else {
+                allText += outText;
+                chapterCount++;
+                if ( chapterCount === epub.flow.length ) {
+                  cb( null, allText );
+                }
+              }
+            });
+          }
+        });
+      }
+    });
+  });
+
+  epub.parse();
+}
+
+module.exports = {
+  types: ['application/epub+zip'],
+  extract: extractText
+};

package.json

@@ -1,6 +1,6 @@
 {
   "name": "textract",
-  "version": "2.3.1",
+  "version": "2.4.0",
   "homepage": "https://github.com/dbashford/textract",
   "author": "David Bashford",
   "description": "Extracting text from files of various type including html, pdf, doc, docx, xls, xlsx, csv, pptx, png, jpg, gif, rtf, text/*, and various open office.",
@@ -43,7 +43,8 @@
     "ots",
     "potx",
     "odg",
-    "otg"
+    "otg",
+    "epub"
   ],
   "dependencies": {
     "mime": "2.2.0",
@@ -58,7 +59,8 @@
     "html-entities": "1.2.0",
     "iconv-lite": "0.4.15",
     "jschardet": "1.4.1",
-    "yauzl": "2.7.0"
+    "yauzl": "2.7.0",
+    "epub2": "1.3.4"
   },
   "devDependencies": {
     "chai": "1.5.0",

test/extract_test.js

@@ -94,6 +94,32 @@ describe( 'textract', function() {
     });
   });
 
+  describe( 'for .epub files', function() {
+    it( 'will extract text from epub files', function( done ) {
+      var docPath = path.join( __dirname, 'files', 'Metamorphosis-jackson.epub' );
+      this.timeout( 5000 );
+      fromFileWithPath( docPath, function( error, text ) {
+        expect( error ).to.be.null;
+        expect( text ).to.be.an( 'string' );
+        expect( text.length ).to.eql( 119329 );
+        expect( text.substring( 3000, 3500 ) ).to.eql( "dboard so that he could lift his head better; found where the itch was, and saw that it was covered with lots of little white spots which he didn't know what to make of; and when he tried to feel the place with one of his legs he drew it quickly back because as soon as he touched it he was overcome by a cold shudder. He slid back into his former position. \"Getting up early all the time\", he thought, \"it makes you stupid. You've got to get enough sleep. Other travelling salesmen live a life of lu" );
+        done();
+      });
+    });
+
+    it( 'will extract text from epub files and preserve line breaks', function( done ) {
+      var docPath = path.join( __dirname, 'files', 'Metamorphosis-jackson.epub' );
+      this.timeout( 5000 );
+      fromFileWithPath( docPath, { preserveLineBreaks: true }, function( error, text ) {
+        expect( error ).to.be.null;
+        expect( text ).to.be.an( 'string' );
+        expect( text.length ).to.eql( 119342 );
+        expect( text.substring( 3000, 3500 ) ).to.eql( "rds the headboard so that he could lift his head better; found where the itch was, and saw that it was covered with lots of little white spots which he didn't know what to make of; and when he tried to feel the place with one of his legs he drew it quickly back because as soon as he touched it he was overcome by a cold shudder.\nHe slid back into his former position. \"Getting up early all the time\", he thought, \"it makes you stupid. You've got to get enough sleep. Other travelling salesmen live a" );
+        done();
+      });
+    });
+  });
+
   describe( 'for .atom files', function() {
     it( 'will extract text from atom files', function( done ) {
       var docPath = path.join( __dirname, 'files', 'atom.atom' );

yarn.lock

@@ -23,6 +23,10 @@ adler-32@:
     exit-on-epipe "~1.0.1"
     printj "~1.1.0"
 
+adm-zip@^0.4.4:
+  version "0.4.11"
+  resolved "https://registry.npmjs.org/adm-zip/-/adm-zip-0.4.11.tgz#2aa54c84c4b01a9d0fb89bb11982a51f13e3d62a"
+
 ajv-keywords@^1.0.0:
   version "1.5.1"
   resolved "https://registry.npmjs.org/ajv-keywords/-/ajv-keywords-1.5.1.tgz#314dd0a4b3368fad3dfcdc54ede6171b886daf3c"
@@ -82,6 +86,10 @@ balanced-match@^1.0.0:
   version "1.0.0"
   resolved "https://registry.npmjs.org/balanced-match/-/balanced-match-1.0.0.tgz#89b4d199ab2bee49de164ea02b89ce462d71b767"
 
+bluebird@^3.5.1:
+  version "3.5.1"
+  resolved "https://registry.npmjs.org/bluebird/-/bluebird-3.5.1.tgz#d9551f9de98f1fcda1e683d17ee91a0602ee2eb9"
+
 boolbase@~1.0.0:
   version "1.0.0"
   resolved "https://registry.npmjs.org/boolbase/-/boolbase-1.0.0.tgz#68dff5fbe60c51eb37725ea9e3ed310dcc1e776e"
@@ -374,6 +382,14 @@ entities@^1.1.1, entities@~1.1.1:
   version "1.1.1"
   resolved "https://registry.npmjs.org/entities/-/entities-1.1.1.tgz#6e5c2d0a5621b5dadaecef80b90edfb5cd7772f0"
 
+[email protected]:
+  version "1.3.4"
+  resolved "https://registry.npmjs.org/epub2/-/epub2-1.3.4.tgz#711fa98f07a99e3dbaba6878cc53b182ca42d436"
+  dependencies:
+    adm-zip "^0.4.4"
+    bluebird "^3.5.1"
+    xml2js "^0.4.4"
+
 error-ex@^1.2.0:
   version "1.3.1"
   resolved "https://registry.npmjs.org/error-ex/-/error-ex-1.3.1.tgz#f855a86ce61adc4e8621c3cda21e7a7612c3a8dc"
@@ -1309,6 +1325,10 @@ safe-buffer@~5.1.0, safe-buffer@~5.1.1:
   version "5.1.1"
   resolved "https://registry.npmjs.org/safe-buffer/-/safe-buffer-5.1.1.tgz#893312af69b2123def71f57889001671eeb2c853"
 
+sax@>=0.6.0:
+  version "1.2.4"
+  resolved "https://registry.npmjs.org/sax/-/sax-1.2.4.tgz#2816234e2378bddc4e5354fab5caa895df7100d9"
+
 "semver@2 || 3 || 4 || 5":
   version "5.5.0"
   resolved "https://registry.npmjs.org/semver/-/semver-5.5.0.tgz#dc4bbc7a6ca9d916dee5d43516f0092b58f7b8ab"
@@ -1516,6 +1536,17 @@ xlsx@~0.7.11:
     jszip "2.4.0"
     ssf "~0.8.1"
 
+xml2js@^0.4.4:
+  version "0.4.19"
+  resolved "https://registry.npmjs.org/xml2js/-/xml2js-0.4.19.tgz#686c20f213209e94abf0d1bcf1efaa291c7827a7"
+  dependencies:
+    sax ">=0.6.0"
+    xmlbuilder "~9.0.1"
+
+xmlbuilder@~9.0.1:
+  version "9.0.7"
+  resolved "https://registry.npmjs.org/xmlbuilder/-/xmlbuilder-9.0.7.tgz#132ee63d2ec5565c557e20f4c22df9aca686b10d"
+
 [email protected]:
   version "0.1.27"
   resolved "https://registry.npmjs.org/xmldom/-/xmldom-0.1.27.tgz#d501f97b3bdb403af8ef9ecc20573187aadac0e9"