Update
February 27, 2014: This article initially only describes the use of PDFBox to parse PDF files. Now it has been extended to include routines that use IFilter and Itextsharp.
This article and the corresponding Visual Studio project have been updated to the current PDFBox version (1.8.4). You can download a complete project from http://www.squarepdf.net/how-to-convert-pdf-to-text-in-net-sample-project/that contains all dependent content (a bit tricky to eliminate dependencies).
How to parse a PDF file
Some of the main ways to extract text from PDF files in. NET are:
Microsoft's IFilter interface and Adobe's IFilter implementation;
Itextsharp;
PDFBox.
Unfortunately, these PDF parsing schemes are not perfect. We will discuss these methods below.
Adobe PDF IFilter
In order to use the IFilter interface to parse PDF files, you need to:
Windows 2000 or later
Adobe Acrobat or Reader 7.0.5+ (or a separate Adobe PDF IFilter [adobe.com])
IFilter COM Encapsulation class [Dotlucene.net]
Sample code:
Using IFilter; ... public static string Extracttextfrompdf (string path) { return defaultparser.extract (path);}
Disadvantages:
Using unreliable COM interop to handle the IFilter interface (and combining IFilter com, Adobe PDF IFilter is particularly troublesome).
Adobe IFilter needs to be installed separately on the target system. If you need to publish an indexed solution to someone else, it can be painful.
Itextsharp
Itextsharp (http://sourceforge.net/projects/itextsharp/) is a. NET output of the Java PDF Operations Library IText (http://itextpdf.com/). It focuses on editing PDFs rather than reading, but it certainly supports extracting text from PDFs (albeit a bit overqualified).
Routines:
Using itextsharp.text.pdf;using iTextSharp.text.pdf.parser; // ... public static string Extracttextfrompdf (string path) { using (pdfreader reader = new Pdfreader (path)) { StringBuilder Text = new StringBuilder (); for (int i = 1; I <= reader. Numberofpages; i++) { text. Append (Pdftextextractor.gettextfrompage (reader, i)); } return text. ToString (); }}
L/C: Member No. 10364982
Disadvantages:
Need a license (if you don't like the AGPL license)
PDFBox
PDFBox is another Java PDF class library. It can also be used in conjunction with the original Java Lucene (see Lucenepdfdocument).
Fortunately, PDFBox has a. NET version developed using Ikvm.net (just visit the PDFBox download page).
Using PDFBox in. NET requires a reference:
IKVM. OpenJDK.Core.dll
IKVM. OpenJDK.SwingAWT.dll
Pdfbox-1.8.4.dll
and copy the following files to the Bin folder:
Commons-logging.dll
Fontbox-1.8.4.dll
IKVM. OpenJDK.Util.dll
IKVM. Runtime.dll
Parsing PDFs with PDFBox is simple:
Using org.apache.pdfbox.pdmodel;using org.apache.pdfbox.util; ... private static string Extracttextfrompdf (string path) { pddocument doc = null; try { doc = pddocument.load (path) pdftextstripper stripper = new Pdftextstripper (); Return Stripper.gettext (DOC); } Finally { if (doc! = null) { doc.close ();}} }
The compiled size adds up to almost 18MB:
IKVM. OpenJDK.Core.dll (4 MB)
IKVM. OpenJDK.SwingAWT.dll (6 MB)
Pdfbox-1.8.4.dll (4 MB)
Commons-logging.dll (KB)
Fontbox-1.8.4.dll (KB)
IKVM. OpenJDK.Util.dll (2 MB)
IKVM. Runtime.dll (1 MB)
Speed can also: Parse U.S. Copyright Act PDF (5.1 MB) file for 13 seconds.
Thank bobrien100 for the suggestions for improvement.
Disadvantages:
IKVM. NET dependency (MB)
Speed (especially ikvm.net start-up time)