From 188e5a8420677c63978020e86fd6f3492fd9bcd4 Mon Sep 17 00:00:00 2001 From: Anjesh Tuladhar Date: Thu, 11 Jun 2015 20:50:31 +0545 Subject: [PATCH] changed the folder structure, and updated text accordingly --- PdfProcessor.py | 7 ++++--- PdfInfo.py => pdftools/PdfInfo.py | 0 PdfSeparate.py => pdftools/PdfSeparate.py | 0 PdfToText.py => pdftools/PdfToText.py | 0 pdftools/__init__.py | 0 run.py | 1 - tests/PdfInfoTest.py | 4 ++-- tests/PdfSeparateTest.py | 2 +- tests/PdfToTextTest.py | 2 +- 9 files changed, 8 insertions(+), 8 deletions(-) rename PdfInfo.py => pdftools/PdfInfo.py (100%) rename PdfSeparate.py => pdftools/PdfSeparate.py (100%) rename PdfToText.py => pdftools/PdfToText.py (100%) create mode 100644 pdftools/__init__.py diff --git a/PdfProcessor.py b/PdfProcessor.py index ca82d6f..de92a1e 100644 --- a/PdfProcessor.py +++ b/PdfProcessor.py @@ -1,9 +1,9 @@ from os import listdir import os.path import json -from PdfInfo import * -from PdfToText import * -from PdfSeparate import * +from pdftools.PdfInfo import * +from pdftools.PdfToText import * +from pdftools.PdfSeparate import * class PDFProcessor: def __init__(self, filePath, outputDir): @@ -18,6 +18,7 @@ def process(self): pdfInfo = PdfInfo(self.filePath) self.totalPages = pdfInfo.getPages() self.fileSize = pdfInfo.getFileSizeInBytes() + self.separatePdfPages() def processToCheckStructured(self): """ diff --git a/PdfInfo.py b/pdftools/PdfInfo.py similarity index 100% rename from PdfInfo.py rename to pdftools/PdfInfo.py diff --git a/PdfSeparate.py b/pdftools/PdfSeparate.py similarity index 100% rename from PdfSeparate.py rename to pdftools/PdfSeparate.py diff --git a/PdfToText.py b/pdftools/PdfToText.py similarity index 100% rename from PdfToText.py rename to pdftools/PdfToText.py diff --git a/pdftools/__init__.py b/pdftools/__init__.py new file mode 100644 index 0000000..e69de29 diff --git a/run.py b/run.py index d89cf29..f4c7175 100644 --- a/run.py +++ b/run.py @@ -11,7 +11,6 @@ pdfProcessor.extractTextFromStructuredDoc() else: pdfProcessor.extractTextFromScannedDoc() -pdfProcessor.separatePdfPages() pdfProcessor.writeStats() diff --git a/tests/PdfInfoTest.py b/tests/PdfInfoTest.py index e921ce6..86d7677 100644 --- a/tests/PdfInfoTest.py +++ b/tests/PdfInfoTest.py @@ -2,8 +2,8 @@ import unittest import sys - -from PdfInfo import * +import os.path +from pdftools.PdfInfo import * class PdfInfoTest(unittest.TestCase): def setUp(self): diff --git a/tests/PdfSeparateTest.py b/tests/PdfSeparateTest.py index b5cea9d..ac47cb1 100644 --- a/tests/PdfSeparateTest.py +++ b/tests/PdfSeparateTest.py @@ -5,7 +5,7 @@ import glob import os -from PdfSeparate import * +from pdftools.PdfSeparate import * class PdfSeparateTest(unittest.TestCase): def setUp(self): diff --git a/tests/PdfToTextTest.py b/tests/PdfToTextTest.py index 3bcc4b1..71fb4ee 100644 --- a/tests/PdfToTextTest.py +++ b/tests/PdfToTextTest.py @@ -5,7 +5,7 @@ import glob import os -from PdfToText import * +from pdftools.PdfToText import * class PdfToTextTest(unittest.TestCase): def setUp(self):