rero · ChristianPereiraIcare · Apr 30, 2018 · May 4, 2018 · May 14, 2018
diff --git a/.gitignore b/.gitignore
@@ -1 +1,3 @@
 *.pyc
+
+\.ropeproject/
diff --git a/Dockerfile b/Dockerfile
@@ -1,37 +1,34 @@
 #
 # Multivio docker build
 #
-FROM debian:jessie
+FROM debian:stretch-slim
 MAINTAINER Johnny Mariéthoz <[email protected]>
 
 # Node.js, bower, less, clean-css, uglify-js, requirejs
 RUN apt-get update
+
 RUN apt-get -qy upgrade --fix-missing --no-install-recommends
 
 # Install dependencies
 RUN apt-get -qy install --fix-missing --no-install-recommends \
-    g++ make git python python-dev python-pip \
-    swig cmake fontconfig libfontconfig1-dev \
-    libjpeg-dev libtiff-dev libopenjpeg-dev \
-    libapache2-mod-wsgi apache2 wget unzip
+    g++ cmake make git python3 python3-dev python3-pip python3-setuptools\
+    fontconfig libfontconfig1-dev zlib1g zlib1g-dev libpng-dev lib32z1-dev\
+    libjpeg-dev libtiff-dev libopenjp2-7-dev \
+    libapache2-mod-wsgi-py3 apache2 wget unzip
+
+RUN pip3 install Cython
 
+RUN pip3 install Pillow
 WORKDIR /code
 
 # Poppler
-RUN git clone git://git.freedesktop.org/git/poppler/poppler 
+RUN git clone git://git.freedesktop.org/git/poppler/poppler
 
 WORKDIR /code/poppler
 
-# Patch poppler > 0.19
-#RUN git checkout -b multivio poppler-0.38.0 \
-#    && perl -pi.bak -e 's/globalParams->getOverprintPreview\(\)/gTrue/g' poppler/SplashOutputDev.h
-
-#poppler 0.18
-RUN git checkout poppler-0.18
-
 RUN mkdir -p /code/poppler/build && cd /code/poppler/build \
 	&& cmake -Wno-dev -D ENABLE_XPDF_HEADERS=True ../ \
-    && make -j 2 install
+    	&& make -j 2 install
 
 # make libpoppler globally available
 RUN ldconfig /usr/local/lib
@@ -41,15 +38,10 @@ COPY . /code/multivio
 WORKDIR /code/multivio
 
 # Basic Python
-RUN pip install --upgrade pip setuptools \
+RUN pip3 install --upgrade pip setuptools \
 	#install multivio
 	&& pip install --global-option=build_ext .
 
-# Multivio client
-
-#RUN adduser --uid 1000 --disabled-password --gecos '' multivio
-#RUN chown -R multivio:multivio /code
-
 # apache
 RUN mkdir -p /var/log/multivio /var/tmp/multivio /var/www/multivio/server \
     && cp tools/multivio_server.py /var/www/multivio/server \
@@ -58,31 +50,16 @@ RUN mkdir -p /var/log/multivio /var/tmp/multivio /var/www/multivio/server \
     && cp tools/multivio.conf /etc/apache2/sites-available/ \
     && a2ensite multivio
 
-# apache script
+# apache sript
 RUN cp scripts/httpd-foreground /usr/local/bin \
     && chmod a+x /usr/local/bin/httpd-foreground
 
-WORKDIR /var/www/multivio/client
-RUN wget http://demo.multivio.org/multivio/client_1.0.0.zip \
-    && unzip client_1.0.0.zip \
-    && mv client_1.0.0/* . \
-    && rm -fr client_1.0.0 client_1.0.0.zip \
-    && chown -R www-data:www-data /var/www/multivio/client
-
-#WORKDIR /
-# Slim down image
-RUN rm -fr /code \
-RUN apt-get clean autoclean \
-   && rm -rf /var/lib/{apt,dpkg}/ \
-   && find /usr/share/doc -depth -type f ! -name copyright -delete \
-   && find /usr/share/doc -empty -delete \
-   && rm -rf /usr/share/man/* /usr/share/groff/* /usr/share/info/* \
-   && rm -rf /tmp/* /var/lib/{cache,log}/ /root/.cache/* \
-   && apt-get -qy remove --purge make git python-dev python-pip swig wget unzip cmake cpp  binutils \
-   && apt-get -qy autoremove
-
-#USER multivio
-#VOLUME ["/code"]
-CMD ["httpd-foreground"]
+#WORKDIR /var/www/multivio/client
+#RUN wget http://demo.multivio.org/multivio/client_1.0.0.zip \
+#    && unzip client_1.0.0.zip \
+#    && mv client_1.0.0/* . \
+#    && rm -fr client_1.0.0 client_1.0.0.zip \
+#cd se    && chown -R www-data:www-data /var/www/multivio/client
 
-#CMD ["rerodoc", "--debug", "run", "-h", "0.0.A0.0"]
+
+CMD ["httpd-foreground"]
diff --git a/multivio/dc_parser.py b/multivio/dc_parser.py
@@ -1,18 +1,18 @@
-#!/usr/bin/env python
+#!/usr/bin/env python3
 # -*- coding: utf-8 -*-
 """Document Parser module for Multivio"""
 
-#==============================================================================
+# ==============================================================================
 #  This file is part of the Multivio software.
 #  Project  : Multivio - https://www.multivio.org/
 #  Copyright: (c) 2009-2011 RERO (http://www.rero.ch/)
 #  License  : See file COPYING
-#==============================================================================
+# ==============================================================================
 
 __copyright__ = "Copyright (c) 2009-2011 RERO"
 __license__ = "GPL V.2"
 
-#---------------------------- Modules ---------------------------------------
+# ---------------------------- Modules ---------------------------------------
 
 # import of standard modules
 import sys
@@ -23,9 +23,10 @@
 from xml.dom.minidom import parseString
 
 # local modules
-from parser import DocumentParser, ParserError
+from multivio.parser import DocumentParser, ParserError
+
+# ----------------------------------- Classes -----------------------------------
 
-#----------------------------------- Classes -----------------------------------
 
 class DublinCoreParser(DocumentParser):
     """To parse PDF document"""
@@ -53,7 +54,7 @@ def _get_record(self):
         self._file_stream.seek(0)
         content_str = self._file_stream.read()
         doc = parseString(content_str)
-        
+
         records = doc.getElementsByTagName('collection')
 
         # get the id number of the first record
@@ -62,7 +63,7 @@ def _get_record(self):
                 "XML/Dublin Core document should contains at lease one record!")
         if len(records) > 1:
             raise ParserError.InvalidDocument(
-                "XML/Dublin Core document should not contains more than "\
+                "XML/Dublin Core document should not contains more than "
                 "one record!")
         return records[0]
 
@@ -71,15 +72,15 @@ def get_metadata(self):
         record = self._get_record()
         metadata = {}
         metadata['title'] = self._get_values_for_labels(record,
-                'title')[0].decode('utf-8')
+                                                        'title')[0].decode('utf-8')
         metadata['creator'] = [v.decode('utf-8') for v in
-            self._get_values_for_labels(record, 'creator')]
+                               self._get_values_for_labels(record, 'creator')]
         metadata['language'] = self._get_values_for_labels(record,
-                'language')[0].decode('utf-8')
-        self.logger.debug("Metadata: %s"% json.dumps(metadata, sort_keys=True, 
-                        indent=4))
+                                                           'language')[0].decode('utf-8')
+        self.logger.debug("Metadata: %s" % json.dumps(metadata, sort_keys=True,
+                                                      indent=4))
         return metadata
-    
+
     def get_physical_structure(self):
         """Get the physical structure of the pdf."""
         phys_struct = []
@@ -90,16 +91,14 @@ def get_physical_structure(self):
                 'url': url,
                 'label': url.split('/')[-1]
             })
-        self.logger.debug("Physical Structure: %s"% json.dumps(phys_struct,
-                sort_keys=True, indent=4))
+        self.logger.debug("Physical Structure: %s" % json.dumps(phys_struct,
+                                                                sort_keys=True, indent=4))
         return phys_struct
 
-
     def _get_values_for_labels(self, record, tag_name):
         """Return the value for a xml label."""
         res = []
         for data_field in record.getElementsByTagNameNS(self._namespace_URI, tag_name):
             if data_field.firstChild is not None:
                 res.append(data_field.firstChild.nodeValue.encode('utf-8'))
         return res
-
diff --git a/multivio/dispatcher_app.py b/multivio/dispatcher_app.py
@@ -1,59 +1,59 @@
-#!/usr/bin/env python
+#!/usr/bin/env python3
 # -*- coding: utf-8 -*-
 """Multivio HTTP requests dispatcher."""
 
-#==============================================================================
+# ==============================================================================
 #  This file is part of the Multivio software.
 #  Project  : Multivio - https://www.multivio.org/
 #  Copyright: (c) 2009-2011 RERO (http://www.rero.ch/)
 #  License  : See file COPYING
-#==============================================================================
+# ==============================================================================
 
 __copyright__ = "Copyright (c) 2009-2011 RERO"
 __license__ = "GPL V.2"
 
 
-#---------------------------- Modules -----------------------------------------
+# ---------------------------- Modules -----------------------------------------
 
 # import of standard modules
 import re
-import sys 
+import sys
 from optparse import OptionParser
 if sys.version_info < (2, 6):
     import simplejson as json
 else:
     import json
 
 # third party modules
-import logger
+import multivio.logger as logger
 import logging
 #import mvo_parser
-import processor_app
-import parser_app
-import version_app
+import multivio.processor_app as processor_app
+import multivio.parser_app as parser_app
+import multivio.version_app as version_app
 
-from web_app import WebApplication
-from webprocessor_app import WebProcessorApp
+from multivio.web_app import WebApplication
+from multivio.webprocessor_app import WebProcessorApp
 from mvo_config import MVOConfig
-from web_app import ApplicationError
+from multivio.web_app import ApplicationError
 
 
 class DispatcherApp(WebApplication):
     """ Dispach http request to several applications given the URI.
-    
+
         This is the entry point of the server application. This class is
         responsible to call applications given the URI of the HTTP request.
     """
 
     def __init__(self):
-        "Simple constructor." 
-        
+        "Simple constructor."
+
         WebApplication.__init__(self)
 
-        #application configuration
+        # application configuration
         self._apps = {}
 
-        #server applications
+        # server applications
         self._apps['.*?/log/post'] = logger.LoggerApp()
         self._apps['.*?/version'] = version_app.VersionApp()
         self._apps['.*?/get.*?'] = \
@@ -64,7 +64,7 @@ def __init__(self):
             WebProcessorApp()
         self.usage = """<br><h1>Welcome to the multivio server.</h1><br>"""
         self.logger = logging.getLogger(MVOConfig.Logger.name+".Dispatcher")
-        
+
     def __call__(self, environ, start_response):
         """Main method to dispatch HTTP requests."""
 
@@ -78,47 +78,48 @@ def __call__(self, environ, start_response):
                 response.extend(["<h3>%s</h3>" % k])
                 self.logger.debug(self._apps[k].usage)
                 response.extend([self._apps[k].usage])
-            return response
+            return response.encode('utf-8')
         for k in self._apps.keys():
+            print(k)  # TODO
             if re.match(k, path):
                 try:
-                    return self._apps[k](environ, start_response)
+                    return self._apps[k](environ, start_response).encode('utf-8')
                 except (ApplicationError.PermissionDenied,
-                    ApplicationError.UnableToRetrieveRemoteDocument,
-                    ApplicationError.UnsupportedFormat,
-                    ApplicationError.InvalidArgument,
-                    ApplicationError.HttpMethodNotAllowed), exception:
+                        ApplicationError.UnableToRetrieveRemoteDocument,
+                        ApplicationError.UnsupportedFormat,
+                        ApplicationError.InvalidArgument,
+                        ApplicationError.HttpMethodNotAllowed) as exception:
                     start_response(exception.http_code, [('content-type',
-                           'application/json')])
+                                                          'application/json')])
                     self.logger.error("Exception: %s occurs with message: %s" %
-                        (type(exception).__name__, str(exception)))
+                                      (type(exception).__name__, str(exception)))
                     result = {
-                        'err_name': type(exception).__name__,
-                        'err_msg' : str(exception)
+                        'err_name1': type(exception).__name__,
+                        'err_msg': str(exception)
                     }
-                    return [json.dumps(result, sort_keys=True, indent=4)]
-                except Exception, exception:
+                    return [(json.dumps(result, sort_keys=True, indent=4)).encode('utf-8')]
+                except Exception as exception:
                     start_response('500 Internal Server Error',
-                            [('content-type', 'application/json')])
+                                   [('content-type', 'application/json')])
                     self.logger.error("Exception: %s occurs with message: %s" %
-                        (type(exception).__name__, str(exception)))
+                                      (type(exception).__name__, str(exception)))
                     result = {
-                        'err_name': type(exception).__name__,
-                        'err_msg' : str(exception)
+                        'err_name2': type(exception).__name__,
+                        'err_msg': str(exception)
                     }
-                    return [json.dumps(result, sort_keys=True, indent=4)]
+                    return [(json.dumps(result, sort_keys=True, indent=4)).encode('utf-8')]
         else:
             self.logger.error("HTTP: 404 for %s" % path)
             start_response('404 File Not Found', [('content-type',
-                            'application/json')])
+                                                   'application/json')])
             result = {
                 'err_name': "FileNotFound",
-                'err_msg' : "File not found"
+                'err_msg': "File not found"
             }
-            return [json.dumps(result, sort_keys=True, indent=4)]
+            return [(json.dumps(result, sort_keys=True, indent=4)).encode('utf-8')]
 
 
-#---------------------------- Main Part ---------------------------------------
+# ---------------------------- Main Part ---------------------------------------
 
 def main():
     """Main function"""
@@ -127,17 +128,17 @@ def main():
 
     parser = OptionParser(usage)
 
-    parser.set_description ("Web app to test the dispatcher.")
+    parser.set_description("Web app to test the dispatcher.")
 
-    parser.add_option ("-v", "--verbose", dest="verbose",
-                       help="Verbose mode",
-                       action="store_true", default=False)
+    parser.add_option("-v", "--verbose", dest="verbose",
+                      help="Verbose mode",
+                      action="store_true", default=False)
 
-    parser.add_option ("-p", "--port", dest="port",
-                       help="Http Port (Default: 4041)",
-                       type="int", default=4041)
+    parser.add_option("-p", "--port", dest="port",
+                      help="Http Port (Default: 4041)",
+                      type="int", default=4041)
 
-    (options, args) = parser.parse_args ()
+    (options, args) = parser.parse_args()
 
     if len(args) != 0:
         parser.error("Error: incorrect number of arguments, try --help")
@@ -146,5 +147,6 @@ def main():
     server = make_server('', options.port, application)
     server.serve_forever()
 
+
 if __name__ == '__main__':
     main()