ajout du parsing au chargement

gargantext_web/views.py

@@ -18,6 +18,8 @@ from django import forms
 
 from collections import defaultdict
 
+from parsing.FileParsers import *
+
 # SOME FUNCTIONS
 
 def query_to_dicts(query_string, *query_args):
@@ -245,14 +247,14 @@ def add_corpus(request):
         form = CorpusForm(request.POST, request.FILES)
         # check whether it's valid:
         if form.is_valid():
-            form.save()
-            # process the data in form.cleaned_data as required
-#            corpus.user = request.user
+            node = form.save()
 #            print(form.cleaned_data['name'])
             
             try:
-                print(type(form.cleaned_data['fichier']))
-                print("here we parse" + str(form.cleaned_data['fichier']))
+                for resource in node.resource.all():
+                    fileparser = PubmedFileParser.PubmedFileParser(file='/var/www/gargantext/media/' + str(resource.file))
+                    fileparser.parse(node)
+
             except Exception as error:
                 print(error)
 

test_parser_ngramextractor.ipynb

 {
  "metadata": {
   "name": "",
-  "signature": "sha256:3345ac991b0346b1dfd82386fdc2a59f39b2de9bf32d03ddfbeb565927cfe7ab"
+  "signature": "sha256:dd0951010ab6bd7a5909f5efbd5f9c0fc791ab47839acdbb492e7d8fbb41f318"
  },
  "nbformat": 3,
  "nbformat_minor": 0,
@@ -19,7 +19,7 @@
      "language": "python",
      "metadata": {},
      "outputs": [],
-     "prompt_number": 1
+     "prompt_number": 10
     },
     {
      "cell_type": "code",
@@ -41,39 +41,106 @@
     {
      "cell_type": "code",
      "collapsed": false,
-     "input": [],
+     "input": [
+      "project_type = NodeType.objects.get(name=\"Project\")"
+     ],
      "language": "python",
      "metadata": {},
      "outputs": [],
-     "prompt_number": 2
+     "prompt_number": 20
+    },
+    {
+     "cell_type": "code",
+     "collapsed": false,
+     "input": [
+      "from node.models import Project\n",
+      "from django.contrib.auth.models import User\n",
+      "user = User.objects.get(username=\"alexandre\")\n",
+      "Project(user=user, type=project_type, name=\"Abeilles\").save()"
+     ],
+     "language": "python",
+     "metadata": {},
+     "outputs": [],
+     "prompt_number": 46
+    },
+    {
+     "cell_type": "code",
+     "collapsed": false,
+     "input": [
+      "node = Node.objects.filter(name=\"Abeilles\", user=user)[0]\n",
+      "print(node.pk)"
+     ],
+     "language": "python",
+     "metadata": {},
+     "outputs": [
+      {
+       "output_type": "stream",
+       "stream": "stdout",
+       "text": [
+        "24\n"
+       ]
+      }
+     ],
+     "prompt_number": 28
+    },
+    {
+     "cell_type": "code",
+     "collapsed": false,
+     "input": [
+      "r = node.children.get(name=\"Pubmed\")"
+     ],
+     "language": "python",
+     "metadata": {},
+     "outputs": [],
+     "prompt_number": 40
     },
     {
      "cell_type": "code",
      "collapsed": false,
      "input": [
-      "node = Node.objects.get(name=\"PubMed corpus\")"
+      "r = r.resource.all()[0]"
      ],
      "language": "python",
      "metadata": {},
      "outputs": [],
-     "prompt_number": 3
+     "prompt_number": 41
+    },
+    {
+     "cell_type": "code",
+     "collapsed": false,
+     "input": [
+      "r.file"
+     ],
+     "language": "python",
+     "metadata": {},
+     "outputs": [
+      {
+       "metadata": {},
+       "output_type": "pyout",
+       "prompt_number": 42,
+       "text": [
+        "<FieldFile: corpora/alexandre/pubmed_BwIXSzN.zip>"
+       ]
+      }
+     ],
+     "prompt_number": 42
     },
     {
      "cell_type": "code",
      "collapsed": false,
      "input": [
-      "fileparser = PubmedFileParser.PubmedFileParser(file='/var/www/gargantext/media/' + node.fichier.name)"
+      "fileparser = PubmedFileParser.PubmedFileParser(file='/var/www/gargantext/media/' + str(r.file))"
      ],
      "language": "python",
      "metadata": {},
      "outputs": [],
-     "prompt_number": 4
+     "prompt_number": 43
     },
     {
      "cell_type": "code",
      "collapsed": false,
      "input": [
-      "fileparser.parse(node)\n"
+      "fileparser.parse(node)"
      ],
      "language": "python",
      "metadata": {},
@@ -179,7 +246,7 @@
       {
        "metadata": {},
        "output_type": "pyout",
-       "prompt_number": 5,
+       "prompt_number": 44,
        "text": [
         "[<Node: Systemic spread and propagation of a plant-pathogenic virus in European honeybees, Apis mellifera.>,\n",
         " <Node: A Causal Analysis of Observed Declines in Managed Honey Bees (Apis mellifera).>,\n",
@@ -265,7 +332,7 @@
        ]
       }
      ],
-     "prompt_number": 5
+     "prompt_number": 44
     },
     {
      "cell_type": "code",
@@ -391,6 +458,28 @@
      ],
      "prompt_number": 26
     },
+    {
+     "cell_type": "code",
+     "collapsed": false,
+     "input": [
+      "Project.objects.all().delete()"
+     ],
+     "language": "python",
+     "metadata": {},
+     "outputs": [],
+     "prompt_number": 18
+    },
+    {
+     "cell_type": "code",
+     "collapsed": false,
+     "input": [
+      "Corpus.objects.all().delete()"
+     ],
+     "language": "python",
+     "metadata": {},
+     "outputs": [],
+     "prompt_number": 45
+    },
     {
      "cell_type": "code",
      "collapsed": false,