Flexibility ++

* Make it easyer to implement different type of processing
* Add example for Pier9
* The generic version now converts pdf to pdf/a before HTML

fs/opt/groomer/functions.py

@@ -2,13 +2,9 @@
 # -*- coding: utf-8 -*-
 import magic
 import os
-import shutil
 import mimetypes
-from twiggy import quickSetup, log
-import shlex
-import subprocess
-import time
 
+from helpers import FileBase, KittenGroomerBase
 
 LIBREOFFICE = '/usr/bin/unoconv'
 GS = '/usr/bin/gs'
@@ -27,21 +23,15 @@ mimes_compressed = ['zip', 'x-rar', 'x-bzip2', 'x-lzip', 'x-lzma', 'x-lzop',
 mimes_data = ['octet-stream']
 
 
-class File(object):
+class File(FileBase):
 
     def __init__(self, src_path, dst_path, main_type, sub_type):
-        self.src_path = src_path
+        super(File, self).__init__(src_path, dst_path)
-        self.dst_path = dst_path
         self.main_type = main_type
         self.sub_type = sub_type
-        self.log_details = {'filepath': self.src_path, 'maintype': self.main_type,
+        self.log_details.update({'maintype': self.main_type, 'subtype': self.sub_type})
-                            'subtype': self.sub_type}
         self.expected_mimetype, self.expected_extensions = self.crosscheck_mime()
         self.is_recursive = False
-        self.log_string = ''
-
-    def add_log_details(self, key, value):
-        self.log_details[key] = value
 
     def crosscheck_mime(self):
         # /usr/share/mime has interesting stuff
@@ -69,39 +59,19 @@ class File(object):
         actual_mimetype = '{}/{}'.format(self.main_type, self.sub_type)
         return actual_mimetype == self.expected_mimetype
 
-    def make_dangerous(self):
-        self.log_details['dangerous'] = True
-        path, filename = os.path.split(self.dst_path)
-        self.dst_path = os.path.join(path, 'DANGEROUS_{}_DANGEROUS'.format(filename))
 
-    def make_unknown(self):
+class KittenGroomer(KittenGroomerBase):
-        self.log_details['unknown'] = True
-        path, filename = os.path.split(self.dst_path)
-        self.dst_path = os.path.join(path, 'UNKNOWN_{}'.format(filename))
 
-    def make_binary(self):
+    def __init__(self, root_src=None, root_dst=None, max_recursive=5):
-        self.log_details['binary'] = True
+        if root_src is None:
-        path, filename = os.path.split(self.dst_path)
+            root_src = os.path.join(os.sep, 'media', 'src')
-        self.dst_path = os.path.join(path, '{}.bin'.format(filename))
+        if root_dst is None:
-
+            root_dst = os.path.join(os.sep, 'media', 'dst')
-
+        super(KittenGroomer, self).__init__(root_src, root_dst)
-class KittenGroomer(object):
-
-    def __init__(self, max_recursive=5):
-        self.src_root_dir = os.path.join(os.sep, 'media', 'src')
-        self.dst_root_dir = os.path.join(os.sep, 'media', 'dst')
-        self.log_root_dir = os.path.join(self.dst_root_dir, 'logs')
-        self.log_processing = os.path.join(self.log_root_dir, 'processing.log')
 
         self.recursive = 0
         self.max_recursive = max_recursive
 
-        # quickSetup(file=self.log_processing)
-        quickSetup()
-        self.log_name = log.name('files')
-
-        self.cur_file = None
-
         subtypes_apps = [
             (mimes_office, self._office_related),
             (mimes_pdf, self._pdf),
@@ -133,30 +103,6 @@ class KittenGroomer(object):
                 to_return[st] = fct
         return to_return
 
-    def _safe_rmtree(self, directory):
-        if os.path.exists(directory):
-            shutil.rmtree(directory)
-
-    def _safe_remove(self, filepath):
-        if os.path.exists(filepath):
-            os.remove(filepath)
-
-    def _safe_mkdir(self, directory):
-        if not os.path.exists(directory):
-            os.makedirs(directory)
-
-    def _safe_copy(self):
-        ''' Create dir if needed '''
-        try:
-            dst_path, filename = os.path.split(self.cur_file.dst_path)
-            self._safe_mkdir(dst_path)
-            shutil.copy(self.cur_file.src_path, self.cur_file.dst_path)
-            return True
-        except Exception as e:
-            # TODO: Logfile
-            print(e)
-            return False
-
     def _list_all_files(self, directory):
         for root, dirs, files in os.walk(directory):
             for filename in files:
@@ -165,16 +111,6 @@ class KittenGroomer(object):
                 maintype, subtype = mimetype.split('/')
                 yield filepath, maintype, subtype
 
-    def _run_process(self, command_line):
-        args = shlex.split(command_line)
-        p = subprocess.Popen(args)
-        while True:
-            code = p.poll()
-            if code is not None:
-                break
-            time.sleep(1)
-        return True
-
     def _print_log(self):
         tmp_log = self.log_name.fields(**self.cur_file.log_details)
         if self.cur_file.log_details.get('dangerous'):
@@ -278,7 +214,7 @@ class KittenGroomer(object):
         extract_command = '{} -p1 x {} -o{} -bd'.format(SEVENZ, self.cur_file.src_path, tmpdir)
         self._run_process(extract_command)
         self.recursive += 1
-        self.processdir(self.cur_file.dst_path, tmpdir)
+        self.processdir(tmpdir, self.cur_file.dst_path)
         self.recursive -= 1
         self._safe_rmtree(tmpdir)
 
@@ -315,11 +251,11 @@ class KittenGroomer(object):
 
     #######################
 
-    def processdir(self, dst_dir=None, src_dir=None):
+    def processdir(self, src_dir=None, dst_dir=None):
-        if dst_dir is None:
-            dst_dir = self.dst_root_dir
         if src_dir is None:
             src_dir = self.src_root_dir
+        if dst_dir is None:
+            dst_dir = self.dst_root_dir
 
         if self.recursive > 0:
             self._print_log()
@@ -344,6 +280,6 @@ class KittenGroomer(object):
                 self._print_log()
 
 if __name__ == '__main__':
-    kg = KittenGroomer()
+    kg = KittenGroomer('/home/raphael/gits/KittenGroomer/tests/content_img_vfat_norm',
-    kg.processdir('/home/raphael/gits/KittenGroomer/tests/content_img_vfat_norm_out',
+                       '/home/raphael/gits/KittenGroomer/tests/content_img_vfat_norm_out')
-                  '/home/raphael/gits/KittenGroomer/tests/content_img_vfat_norm')
+    kg.processdir()

fs/opt/groomer/functions_pier9.py

@@ -0,0 +1,62 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+import os
+
+from helpers import FileBase, KittenGroomerBase
+
+
+printers = ['.STL', '.obj']
+cnc = ['.nc', '.tap', '.gcode', '.dxf', '.stl', '.obj', '.iges', '.igs',
+       '.vrml', '.vrl', '.thing', '.step', '.stp', '.x3d']
+shopbot = ['.ai', '.svg', '.dxf', '.dwg', '.eps']
+omax = ['.ai', '.svg', '.dxf', '.dwg', '.eps', '.omx', '.obj']
+epilog_laser = ['.ai', '.svg', '.dxf', '.dwg', '.eps']
+metabeam = ['.dxf']
+up = ['.upp', '.up3', '.stl', '.obj']
+
+
+class FilePier9(FileBase):
+
+    def __init__(self, src_path, dst_path):
+        super(FilePier9, self).__init__(src_path, dst_path)
+        a, self.extension = os.path.splitext(self.src_path)
+
+
+class KittenGroomerPier9(KittenGroomerBase):
+
+    def __init__(self, root_src=None, root_dst=None):
+
+        if root_src is None:
+            root_src = os.path.join(os.sep, 'media', 'src')
+        if root_dst is None:
+            root_dst = os.path.join(os.sep, 'media', 'dst')
+        super(KittenGroomerPier9, self).__init__(root_src, root_dst)
+
+        # The initial version will accept all the file extension for all the machines.
+        self.authorized_extensions = printers + cnc + shopbot + omax + epilog_laser + metabeam + up
+
+    def _print_log(self):
+        tmp_log = self.log_name.fields(**self.cur_file.log_details)
+        if not self.cur_file.log_details.get('valid'):
+            tmp_log.warning(self.cur_file.log_string)
+        else:
+            tmp_log.debug(self.cur_file.log_string)
+
+    def processdir(self):
+
+        for srcpath in self._list_all_files(self.src_root_dir):
+            self.log_name.info('Processing {}', srcpath.replace(self.src_root_dir + '/', ''))
+            self.cur_file = FilePier9(srcpath, srcpath.replace(self.src_root_dir, self.dst_root_dir))
+            if self.cur_file.extension in self.authorized_extensions:
+                self.cur_file.add_log_details('valid', True)
+                self.cur_file.log_string = 'Expected extension: ' + self.cur_file.extension
+                self._safe_copy()
+            else:
+                self.cur_file.log_string = 'Bad extension: ' + self.cur_file.extension
+            self._print_log()
+
+
+if __name__ == '__main__':
+    kg = KittenGroomerPier9('/home/raphael/gits/KittenGroomer/tests/content_img_vfat_norm',
+                            '/home/raphael/gits/KittenGroomer/tests/content_img_vfat_norm_out')
+    kg.processdir()

fs/opt/groomer/helpers.py

@@ -0,0 +1,110 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+import os
+import shutil
+from twiggy import quickSetup, log
+import shlex
+import subprocess
+import time
+
+
+class KittenGroomerError(Exception):
+    def __init__(self, message):
+        super(KittenGroomerError, self).__init__(message)
+        self.message = message
+
+
+class ImplementationRequired(KittenGroomerError):
+    pass
+
+
+class FileBase(object):
+
+    def __init__(self, src_path, dst_path):
+        self.src_path = src_path
+        self.dst_path = dst_path
+        self.log_details = {'filepath': self.src_path}
+        self.log_string = ''
+
+    def add_log_details(self, key, value):
+        self.log_details[key] = value
+
+    def make_dangerous(self):
+        self.log_details['dangerous'] = True
+        path, filename = os.path.split(self.dst_path)
+        self.dst_path = os.path.join(path, 'DANGEROUS_{}_DANGEROUS'.format(filename))
+
+    def make_unknown(self):
+        self.log_details['unknown'] = True
+        path, filename = os.path.split(self.dst_path)
+        self.dst_path = os.path.join(path, 'UNKNOWN_{}'.format(filename))
+
+    def make_binary(self):
+        self.log_details['binary'] = True
+        path, filename = os.path.split(self.dst_path)
+        self.dst_path = os.path.join(path, '{}.bin'.format(filename))
+
+
+class KittenGroomerBase(object):
+
+    def __init__(self, root_src, root_dst):
+        self.src_root_dir = root_src
+        self.dst_root_dir = root_dst
+        self.log_root_dir = os.path.join(self.dst_root_dir, 'logs')
+        self.log_processing = os.path.join(self.log_root_dir, 'processing.log')
+
+        # quickSetup(file=self.log_processing)
+        quickSetup()
+        self.log_name = log.name('files')
+
+        self.cur_file = None
+
+    # ##### Helpers #####
+    def _safe_rmtree(self, directory):
+        if os.path.exists(directory):
+            shutil.rmtree(directory)
+
+    def _safe_remove(self, filepath):
+        if os.path.exists(filepath):
+            os.remove(filepath)
+
+    def _safe_mkdir(self, directory):
+        if not os.path.exists(directory):
+            os.makedirs(directory)
+
+    def _safe_copy(self):
+        ''' Create dir if needed '''
+        try:
+            dst_path, filename = os.path.split(self.cur_file.dst_path)
+            self._safe_mkdir(dst_path)
+            shutil.copy(self.cur_file.src_path, self.cur_file.dst_path)
+            return True
+        except Exception as e:
+            # TODO: Logfile
+            print(e)
+            return False
+
+    def _list_all_files(self, directory):
+        for root, dirs, files in os.walk(directory):
+            for filename in files:
+                filepath = os.path.join(root, filename)
+                yield filepath
+
+    def _run_process(self, command_line):
+        args = shlex.split(command_line)
+        p = subprocess.Popen(args)
+        while True:
+            code = p.poll()
+            if code is not None:
+                break
+            time.sleep(1)
+        return True
+
+    def _print_log(self):
+        # Not implemented
+        pass
+
+    #######################
+
+    def processdir(self, src_dir=None, dst_dir=None):
+        raise ImplementationRequired('You have to implement the result processdir.')