Merge branch 'master' of stigma.nowoczesnapolska.org.pl:platforma
[redakcja.git] / apps / api / handlers / library_handlers.py
index 86aef21..488c2d4 100644 (file)
@@ -1,4 +1,6 @@
 # -*- encoding: utf-8 -*-
+import os.path
+import logging
 
 __author__= "Ɓukasz Rekucki"
 __date__ = "$2009-09-25 15:49:50$"
@@ -6,25 +8,37 @@ __doc__ = "Module documentation."
 
 from piston.handler import BaseHandler, AnonymousBaseHandler
 
+import re
+from datetime import date
+
+from django.core.urlresolvers import reverse
+from django.utils import simplejson as json
+from django.db import IntegrityError
 
 import librarian
 import librarian.html
+from librarian import dcparser, parser
+
+from wlrepo import *
+from explorer.models import PullRequest, GalleryForDocument
+
+# internal imports
 import api.forms as forms
-from datetime import date
+import api.response as response
+from api.utils import validate_form, hglibrary, natural_order
+from api.models import PartCache
 
-from django.core.urlresolvers import reverse
+#
+import settings
 
-from wlrepo import RevisionNotFound, LibraryException, DocumentAlreadyExists
-from librarian import dcparser
 
-import api.response as response
-from api.utils import validate_form, hglibrary
+log = logging.getLogger('platforma.api')
 
-from explorer.models import PullRequest
 
 #
 # Document List Handlers
 #
+# TODO: security check
 class BasicLibraryHandler(AnonymousBaseHandler):
     allowed_methods = ('GET',)
 
@@ -38,19 +52,50 @@ class BasicLibraryHandler(AnonymousBaseHandler):
         return {'documents' : document_list}
         
 
+#
+# This handler controlls the document collection
+#
 class LibraryHandler(BaseHandler):
     allowed_methods = ('GET', 'POST')
     anonymous = BasicLibraryHandler
 
+
     @hglibrary
     def read(self, request, lib):
         """Return the list of documents."""
 
-        document_list = [{
-            'url': reverse('document_view', args=[docid]),
-            'name': docid } for docid in lib.documents() ]
-
-        return {'documents' : document_list }        
+        documents = {}
+        
+        for docid in lib.documents():            
+            documents[docid] = {
+                'url': reverse('document_view', args=[docid]),
+                'name': docid,
+                'parts': []
+            }
+
+        parts = PartCache.objects.defer('part_id')\
+            .values_list('part_id', 'document_id').distinct()
+       
+        document_tree = dict(documents)
+
+        for part, docid in parts:
+            # this way, we won't display broken links
+            if not documents.has_key(part):
+                log.info("NOT FOUND: %s", part)
+                continue
+
+            parent = documents[docid]
+            child = documents[part]
+
+            # not top-level anymore
+            document_tree.pop(part)
+            parent['parts'].append(child)
+        
+        for doc in documents.itervalues():
+            doc['parts'].sort(key=natural_order(lambda d: d['name']))
+            
+        return {'documents': sorted(document_tree.itervalues(),
+            key=natural_order(lambda d: d['name']) ) }
 
     @validate_form(forms.DocumentUploadForm, 'POST')
     @hglibrary
@@ -62,6 +107,9 @@ class LibraryHandler(BaseHandler):
         else:            
             data = request.FILES['ocr_file'].read().decode('utf-8')
 
+        if data is None:
+            return response.BadRequest().django_response('You must pass ocr_data or ocr_file.')
+
         if form.cleaned_data['generate_dc']:
             data = librarian.wrap_text(data, unicode(date.today()))
 
@@ -70,7 +118,7 @@ class LibraryHandler(BaseHandler):
         try:
             lock = lib.lock()            
             try:
-                print "DOCID", docid                
+                log.info("DOCID %s", docid)
                 doc = lib.document_create(docid)
                 # document created, but no content yet
 
@@ -78,9 +126,10 @@ class LibraryHandler(BaseHandler):
                     doc = doc.quickwrite('xml', data.encode('utf-8'),
                         '$AUTO$ XML data uploaded.', user=request.user.username)
                 except Exception,e:
+                    import traceback
                     # rollback branch creation
                     lib._rollback()
-                    raise LibraryException("Exception occured:" + repr(e))
+                    raise LibraryException(traceback.format_exc())
 
                 url = reverse('document_view', args=[doc.id])
 
@@ -93,12 +142,16 @@ class LibraryHandler(BaseHandler):
             finally:
                 lock.release()
         except LibraryException, e:
-            return response.InternalError().django_response(\
-                {'exception': repr(e) })                
+            import traceback
+            return response.InternalError().django_response({
+                "reason": traceback.format_exc()
+            })
         except DocumentAlreadyExists:
             # Document is already there
-            return response.EntityConflict().django_response(\
-                {"reason": "Document %s already exists." % docid})
+            return response.EntityConflict().django_response({
+                "reason": "already-exists",
+                "message": "Document already exists." % docid
+            })
 
 #
 # Document Handlers
@@ -115,9 +168,9 @@ class BasicDocumentHandler(AnonymousBaseHandler):
 
         result = {
             'name': doc.id,
-            'html_url': reverse('dochtml_view', args=[doc.id,doc.revision]),
-            'text_url': reverse('doctext_view', args=[doc.id,doc.revision]),
-            'dc_url': reverse('docdc_view', args=[doc.id,doc.revision]),
+            'html_url': reverse('dochtml_view', args=[doc.id]),
+            'text_url': reverse('doctext_view', args=[doc.id]),
+            'dc_url': reverse('docdc_view', args=[doc.id]),
             'public_revision': doc.revision,
         }
 
@@ -133,22 +186,29 @@ class DocumentHandler(BaseHandler):
     @hglibrary
     def read(self, request, docid, lib):
         """Read document's meta data"""       
+        log.info(u"Read %s (%s)" % (docid, type(docid)) )
         try:
             doc = lib.document(docid)
             udoc = doc.take(request.user.username)
-        except RevisionNotFound:
-            return request.EnityNotFound().django_response()
+        except RevisionNotFound, e:
+            return response.EntityNotFound().django_response({
+                'exception': type(e), 'message': e.message,
+                'docid': docid })
 
         # is_shared = udoc.ancestorof(doc)
         # is_uptodate = is_shared or shared.ancestorof(document)
 
         result = {
             'name': udoc.id,
-            'html_url': reverse('dochtml_view', args=[udoc.id,udoc.revision]),
-            'text_url': reverse('doctext_view', args=[udoc.id,udoc.revision]),
-            'dc_url': reverse('docdc_view', args=[udoc.id,udoc.revision]),
+            'html_url': reverse('dochtml_view', args=[udoc.id]),
+            'text_url': reverse('doctext_view', args=[udoc.id]),
+            'dc_url': reverse('docdc_view', args=[udoc.id]),
+            'gallery_url': reverse('docgallery_view', args=[udoc.id]),
+            'merge_url': reverse('docmerge_view', args=[udoc.id]),
             'user_revision': udoc.revision,
-            'public_revision': doc.revision,            
+            'user_timestamp': udoc.revision.timestamp,
+            'public_revision': doc.revision,
+            'public_timestamp': doc.revision.timestamp,
         }       
 
         return result
@@ -160,52 +220,137 @@ class DocumentHandler(BaseHandler):
 #
 #
 #
-
 class DocumentHTMLHandler(BaseHandler):
-    allowed_methods = ('GET', 'PUT')
+    allowed_methods = ('GET')
 
     @hglibrary
-    def read(self, request, docid, revision, lib):
+    def read(self, request, docid, lib, stylesheet='partial'):
         """Read document as html text"""
         try:
+            revision = request.GET.get('revision', 'latest')
+
             if revision == 'latest':
                 document = lib.document(docid)
             else:
                 document = lib.document_for_rev(revision)
 
-            return librarian.html.transform(document.data('xml'))
-        except RevisionNotFound:
-            return response.EntityNotFound().django_response()
+            if document.id != docid:
+                return response.BadRequest().django_response({'reason': 'name-mismatch',
+                    'message': 'Provided revision refers, to document "%s", but provided "%s"' % (document.id, docid) })
+
+            return librarian.html.transform(document.data('xml'), is_file=False, \
+                parse_dublincore=False, stylesheet=stylesheet,\
+                options={
+                    "with-paths": 'boolean(1)',                    
+                })
+                
+        except (EntryNotFound, RevisionNotFound), e:
+            return response.EntityNotFound().django_response({
+                'reason': 'not-found', 'message': e.message})
+        except librarian.ParseError, e:
+            return response.InternalError().django_response({
+                'reason': 'xml-parse-error', 'message': e.message })
+
+#
+# Image Gallery
+#
+
+class DocumentGalleryHandler(BaseHandler):
+    allowed_methods = ('GET')
+    
+    
+    def read(self, request, docid):
+        """Read meta-data about scans for gallery of this document."""
+        galleries = []
+        from urllib import quote
+
+        for assoc in GalleryForDocument.objects.filter(document=docid):
+            dirpath = os.path.join(settings.MEDIA_ROOT, assoc.subpath)
+
+            if not os.path.isdir(dirpath):
+                log.warn(u"[WARNING]: missing gallery %s", dirpath)
+                continue
+
+            gallery = {'name': assoc.name, 'pages': []}
+            
+            for file in os.listdir(dirpath):
+                if not isinstance(file, unicode):
+                    log.warn(u"File %r is gallery %r is not unicode. Ommiting."\
+                        % (file, dirpath) )
+                    continue
+                               
+                name, ext = os.path.splitext(os.path.basename(file))
+
+                if ext.lower() not in [u'.png', u'.jpeg', u'.jpg']:
+                    log.info(u"Ignoring: %s %s", name, ext)
+                    continue
+
+                url = settings.MEDIA_URL + assoc.subpath + u'/' + file;
+                gallery['pages'].append( quote(url.encode('utf-8')) )
+
+            gallery['pages'].sort()
+            galleries.append(gallery)
+
+        return galleries                      
 
 #
 # Document Text View
 #
+
+XINCLUDE_REGEXP = r"""<(?:\w+:)?include\s+[^>]*?href=("|')wlrepo://(?P<link>[^\1]+?)\1\s*[^>]*?>"""
+#
+#
+#
+
 class DocumentTextHandler(BaseHandler):
-    allowed_methods = ('GET', 'PUT')
+    allowed_methods = ('GET', 'POST')
 
     @hglibrary
-    def read(self, request, docid, revision, lib):
-        """Read document as raw text"""               
+    def read(self, request, docid, lib):
+        """Read document as raw text"""
+        revision = request.GET.get('revision', 'latest')
+        part = request.GET.get('part', False)
+        
         try:
             if revision == 'latest':
                 document = lib.document(docid)
             else:
                 document = lib.document_for_rev(revision)
+
+            if document.id != docid:
+                return response.BadRequest().django_response({'reason': 'name-mismatch',
+                    'message': 'Provided revision is not valid for this document'})
             
             # TODO: some finer-grained access control
-            return document.data('xml')
-        except RevisionNotFound:
-            return response.EntityNotFound().django_response()
+            if part is False:
+                # we're done :)
+                return document.data('xml')
+            else:
+                xdoc = parser.WLDocument.from_string(document.data('xml'),\
+                    parse_dublincore=False)
+                ptext = xdoc.part_as_text(part)
+
+                if ptext is None:
+                    return response.EntityNotFound().django_response({
+                      'reason': 'no-part-in-document'                     
+                    })
+
+                return ptext
+        except librarian.ParseError:
+            return response.EntityNotFound().django_response({
+                'reason': 'invalid-document-state',
+                'exception': type(e), 'message': e.message
+            })
+        except (EntryNotFound, RevisionNotFound), e:
+            return response.EntityNotFound().django_response({
+                'reason': 'not-found',
+                'exception': type(e), 'message': e.message
+            })   
 
     @hglibrary
-    def update(self, request, docid, revision, lib):
+    def create(self, request, docid, lib):
         try:
-            data = request.PUT['contents']            
-
-            if request.PUT.has_key('message'):
-                msg = u"$USER$ " + request.PUT['message']
-            else:
-                msg = u"$AUTO$ XML content update."
+            revision = request.POST['revision']
 
             current = lib.document(docid, request.user.username)
             orig = lib.document_for_rev(revision)
@@ -216,21 +361,85 @@ class DocumentTextHandler(BaseHandler):
                         "provided_revision": orig.revision,
                         "latest_revision": current.revision })
 
-            ndoc = current.quickwrite('xml', data, msg)
+            if request.POST.has_key('message'):
+                msg = u"$USER$ " + request.POST['message']
+            else:
+                msg = u"$AUTO$ XML content update."
+
+            if request.POST.has_key('contents'):
+                data = request.POST['contents']
+            else:
+                if not request.POST.has_key('chunks'):
+                    # bad request
+                    return response.BadRequest().django_response({'reason': 'invalid-arguments',
+                        'message': 'No contents nor chunks specified.'})
+
+                    # TODO: validate
+                parts = json.loads(request.POST['chunks'])                    
+                xdoc = parser.WLDocument.from_string(current.data('xml'))
+                   
+                errors = xdoc.merge_chunks(parts)
+
+                if len(errors):
+                    return response.EntityConflict().django_response({
+                            "reason": "invalid-chunks",
+                            "message": "Unable to merge following parts into the document: %s " % ",".join(errors)
+                    })
+
+                data = xdoc.serialize()
+
+            # try to find any Xinclude tags
+            includes = [m.groupdict()['link'] for m in (re.finditer(\
+                XINCLUDE_REGEXP, data, flags=re.UNICODE) or []) ]
+
+            log.info("INCLUDES: %s", includes)
+
+            # TODO: provide useful routines to make this simpler
+            def xml_update_action(lib, resolve):
+                try:
+                    f = lib._fileopen(resolve('parts'), 'r')
+                    stored_includes = json.loads(f.read())
+                    f.close()
+                except:
+                    stored_includes = []
+                
+                if stored_includes != includes:
+                    f = lib._fileopen(resolve('parts'), 'w+')
+                    f.write(json.dumps(includes))
+                    f.close()
+
+                    lib._fileadd(resolve('parts'))
+
+                    # update the parts cache
+                    PartCache.update_cache(docid, current.owner,\
+                        stored_includes, includes)
+
+                # now that the parts are ok, write xml
+                f = lib._fileopen(resolve('xml'), 'w+')
+                f.write(data.encode('utf-8'))
+                f.close()
+
+            ndoc = None
+            ndoc = current.invoke_and_commit(\
+                xml_update_action, lambda d: (msg, current.owner) )
 
             try:
                 # return the new revision number
-                return {
+                return response.SuccessAllOk().django_response({
                     "document": ndoc.id,
                     "subview": "xml",
                     "previous_revision": current.revision,
-                    "updated_revision": ndoc.revision
-                }
+                    "revision": ndoc.revision,
+                    'timestamp': ndoc.revision.timestamp,
+                    "url": reverse("doctext_view", args=[ndoc.id])
+                })
             except Exception, e:
-                lib._rollback()
+                if ndoc: lib._rollback()
                 raise e        
         except RevisionNotFound, e:
-            return response.EntityNotFound().django_response(e)
+            return response.EntityNotFound(mimetype="text/plain").\
+                django_response(e.message)
+
 
 #
 # Dublin Core handlers
@@ -238,27 +447,37 @@ class DocumentTextHandler(BaseHandler):
 # @requires librarian
 #
 class DocumentDublinCoreHandler(BaseHandler):
-    allowed_methods = ('GET', 'PUT')
+    allowed_methods = ('GET', 'POST')
 
     @hglibrary
-    def read(self, request, docid, revision, lib):
+    def read(self, request, docid, lib):
         """Read document as raw text"""        
         try:
+            revision = request.GET.get('revision', 'latest')
+
             if revision == 'latest':
                 doc = lib.document(docid)
             else:
                 doc = lib.document_for_rev(revision)
+
+
+            if document.id != docid:
+                return response.BadRequest().django_response({'reason': 'name-mismatch',
+                    'message': 'Provided revision is not valid for this document'})
             
             bookinfo = dcparser.BookInfo.from_string(doc.data('xml'))
             return bookinfo.serialize()
-        except RevisionNotFound:
-            return response.EntityNotFound().django_response()
+        except (EntryNotFound, RevisionNotFound), e:
+            return response.EntityNotFound().django_response({
+                'exception': type(e), 'message': e.message})
 
     @hglibrary
-    def update(self, request, docid, revision, lib):
+    def create(self, request, docid, lib):
         try:
-            bi_json = request.PUT['contents']            
-            if request.PUT.has_key('message'):
+            bi_json = request.POST['contents']
+            revision = request.POST['revision']
+            
+            if request.POST.has_key('message'):
                 msg = u"$USER$ " + request.PUT['message']
             else:
                 msg = u"$AUTO$ Dublin core update."
@@ -286,10 +505,12 @@ class DocumentDublinCoreHandler(BaseHandler):
                     "document": ndoc.id,
                     "subview": "dc",
                     "previous_revision": current.revision,
-                    "updated_revision": ndoc.revision
+                    "revision": ndoc.revision,
+                    'timestamp': ndoc.revision.timestamp,
+                    "url": reverse("docdc_view", args=[ndoc.id])
                 }
             except Exception, e:
-                lib._rollback()
+                if ndoc: lib._rollback()
                 raise e
         except RevisionNotFound:
             return response.EntityNotFound().django_response()
@@ -335,41 +556,50 @@ class MergeHandler(BaseHandler):
                     "provided": target_rev,
                     "latest": udoc.revision })
 
-        if not request.user.has_perm('explorer.book.can_share'):
-            # User is not permitted to make a merge, right away
-            # So we instead create a pull request in the database
-            prq = PullRequest(
-                comitter=request.user,
-                document=docid,
-                source_revision = str(udoc.revision),
-                status="N",
-                comment = form.cleaned_data['comment'] or '$AUTO$ Document shared.'
-            )
-
-            prq.save()
-            return response.RequestAccepted().django_response(\
-                ticket_status=prq.status, \
-                ticket_uri=reverse("pullrequest_view", args=[prq.id]) )
-
-        if form.cleanded_data['type'] == 'update':
+        if form.cleaned_data['type'] == 'update':
             # update is always performed from the file branch
             # to the user branch
             success, changed = udoc.update(request.user.username)
 
-        if form.cleanded_data['type'] == 'share':
-            success, changed = udoc.share(form.cleaned_data['comment'])
+        if form.cleaned_data['type'] == 'share':        
+            if not request.user.has_perm('explorer.document.can_share'):
+                # User is not permitted to make a merge, right away
+                # So we instead create a pull request in the database
+                try:
+                    prq, created = PullRequest.objects.get_or_create(
+                        source_revision = str(udoc.revision),
+                        defaults = {
+                            'comitter': request.user,
+                            'document': docid,
+                            'status': "N",
+                            'comment': form.cleaned_data['message'] or '$AUTO$ Document shared.',
+                        }
+                    )
+
+                    return response.RequestAccepted().django_response(\
+                        ticket_status=prq.status, \
+                        ticket_uri=reverse("pullrequest_view", args=[prq.id]) )
+                except IntegrityError:
+                    return response.EntityConflict().django_response({
+                        'reason': 'request-already-exist'
+                    })
+            else:
+                success, changed = udoc.share(form.cleaned_data['message'])
 
         if not success:
-            return response.EntityConflict().django_response()
+            return response.EntityConflict().django_response({
+                'reason': 'merge-failure',
+            })
 
         if not changed:
             return response.SuccessNoContent().django_response()
 
-        new_udoc = udoc.latest()
+        nudoc = udoc.latest()
 
         return response.SuccessAllOk().django_response({
-            "name": udoc.id,
+            "name": nudoc.id,
             "parent_user_resivion": udoc.revision,
             "parent_revision": doc.revision,
-            "revision": udoc.revision,
-        })
\ No newline at end of file
+            "revision": nudoc.revision,
+            'timestamp': nudoc.revision.timestamp,
+        })