new frutch + spirale

[iramuteq] / parse_factiva_txt.py

diff --git a/parse_factiva_txt.py b/parse_factiva_txt.py
index 9cb2af2..fd856f4 100644 (file)
--- a/parse_factiva_txt.py
+++ b/parse_factiva_txt.py
@@ -2,10 +2,11 @@
 # -*- coding: utf-8 -*-
 #Author: Pierre Ratinaud
 #Copyright (c) 2012-2013 Pierre Ratinaud
 # -*- coding: utf-8 -*-
 #Author: Pierre Ratinaud
 #Copyright (c) 2012-2013 Pierre Ratinaud

-#Lisense: GNU/GPL
+#License: GNU/GPL

 
 import os
 import codecs
 
 import os
 import codecs

+import re

 
 
 #txtdir = 'dev/factiva_txt'
 
 
 #txtdir = 'dev/factiva_txt'


@@ -13,6 +14,31 @@ import codecs
 #encodage_in = 'utf8'
 #encodage_out = 'utf8'
 
 #encodage_in = 'utf8'
 #encodage_out = 'utf8'
 

+mois = {u'janvier' : '01', 
+        u'février' : '02',
+        u'mars' : '03',
+        u'avril' : '04', 
+        u'mai' : '05',
+        u'juin' : '06',
+        u'juillet' : '07',
+        u'août' : '08',
+        u'septembre' : '09',
+        u'octobre' : '10',
+        u'novembre' : '11',
+        u'décembre' : '12', 
+        u'january' : '01',
+        u'february': '02',
+        u'march' : '03',
+        u'april': '04',
+        u'may': '05',
+        u'june' : '06',
+        u'july': '07',
+        u'august': '08',
+        u'september' : '09',
+        u'october': '10',
+        u'november': '11',
+        u'december': '12'}
+

 
 def parsetxtpaste(txt):
     """
 
 def parsetxtpaste(txt):
     """


@@ -25,19 +51,27 @@ def parsetxtpaste(txt):
     keepline = False
     ucis = []
     for line in txt : 
     keepline = False
     ucis = []
     for line in txt : 

-        if line.startswith('Article') :
+        if line.startswith(u'Article') :

             lp = line.split()
             if len(lp) > 2  :
             lp = line.split()
             if len(lp) > 2  :

-                if lp[2] == 'Article' :
+                if lp[2] == u'Article' or lp[2] == u'Next' or lp[2] == u'Previous':

                     ucis.append([[u'****'],''])
                     keepline = False
         if line.startswith('SN ') : #source
                     ucis.append([[u'****'],''])
                     keepline = False
         if line.startswith('SN ') : #source

-            source = '*source_' + line[4:].replace(' ','').replace('\'','').replace(u'´','').replace(u'’','').replace('-','').lower()
+            jsource = re.sub(u'[\'" !\.?;,:\+\-°&]', '', line[4:])
+            source = u'_'.join([u'*source', jsource]).lower()
+            #source = '*source_' + line[4:].replace(' ','').replace('\'','').replace(u'´','').replace(u'’','').replace('-','').lower()

             ucis[-1][0].append(source)
         elif line.startswith('PD ') : #date
             ucis[-1][0].append(source)
         elif line.startswith('PD ') : #date

-            mois_annee = '*ma_' + line[4:].split(' ')[1] + line[4:].split(' ')[2]
-            ucis[-1][0].append(mois_annee)
-            annee = u'*annee_' + line[4:].split(' ')[2]
+            datemois = line[4:].split(' ')[1].lower()
+            datemois = mois.get(datemois, datemois)
+            dateannee = line[4:].split(' ')[2]
+            datejour = '%02d' % int(line[4:].split(' ')[0])
+            am = '_'.join([u'*am', dateannee, datemois])
+            amj = '_'.join([u'*amj', dateannee, datemois, datejour])
+            ucis[-1][0].append(am)
+            ucis[-1][0].append(amj)
+            annee = '_'.join([u'*annee', dateannee])

             ucis[-1][0].append(annee)
         elif line.strip() in no : #fin
             keepline = False
             ucis[-1][0].append(annee)
         elif line.strip() in no : #fin
             keepline = False


@@ -48,7 +82,7 @@ def parsetxtpaste(txt):
         else :
             pass
         if keepline and line.strip() not in ['LP', 'TD', ''] :
         else :
             pass
         if keepline and line.strip() not in ['LP', 'TD', ''] :

-            ucis[-1][1] = '\n'.join([ucis[-1][1],line])
+            ucis[-1][1] = '\n'.join([ucis[-1][1],line.replace(u'*', ' ')])

     return ucis
 
 
     return ucis
 
 


@@ -56,18 +90,25 @@ def print_ucis(ucis, ofile, encodage) :
     #elimination des articles vides
     ucis = [uci for uci in ucis if uci[1].strip() != '']
     toprint = '\n\n'.join(['\n'.join([' '.join(uci[0]),uci[1]]) for uci in ucis])
     #elimination des articles vides
     ucis = [uci for uci in ucis if uci[1].strip() != '']
     toprint = '\n\n'.join(['\n'.join([' '.join(uci[0]),uci[1]]) for uci in ucis])

-    ofile.write(toprint.encode(encodage))
+    ofile.write(toprint.encode(encodage, errors='replace') + '\n')

 
 class ParseFactivaPaste :
     def __init__(self, txtdir, fileout, encodage_in, encodage_out) :
 
 class ParseFactivaPaste :
     def __init__(self, txtdir, fileout, encodage_in, encodage_out) :

-        files = os.listdir(txtdir) 
+        files = []
+        for root, subfolders, subfiles in os.walk(txtdir) :
+            nf = [os.path.join(root, f) for f in subfiles if f.split('.')[-1] == 'txt']
+            nf.sort()
+            files += nf
+        tot = 0

         with open(fileout,'w') as outf : 
             for f in files : 
         with open(fileout,'w') as outf : 
             for f in files : 

-                f= os.path.join(txtdir, f) 
+                print f

                 with codecs.open(f, 'rU', encodage_in) as infile : 
                     content = infile.read() 
                 ucis = parsetxtpaste(content)
                 print_ucis(ucis, outf, encodage_out)
                 with codecs.open(f, 'rU', encodage_in) as infile : 
                     content = infile.read() 
                 ucis = parsetxtpaste(content)
                 print_ucis(ucis, outf, encodage_out)

+                tot += len(ucis)
+                print 'ok', len(ucis), 'articles', ' - total : ', tot

 
 #for dat in ['2001','2002','2003','2004', '2005','2006','2007','2008','2009','2010','2011'] :
 #    path = os.path.join(txtdir,dat)
 
 #for dat in ['2001','2002','2003','2004', '2005','2006','2007','2008','2009','2010','2011'] :
 #    path = os.path.join(txtdir,dat)