apps/api/handlers/library_handlers.py

   1 # -*- encoding: utf-8 -*-
   2 import os.path
   3
   4 import logging
   5 log = logging.getLogger('platforma.api.library')
   6
   7 __author__= "Łukasz Rekucki"
   8 __date__ = "$2009-09-25 15:49:50$"
   9 __doc__ = "Module documentation."
  10
  11 from piston.handler import BaseHandler, AnonymousBaseHandler
  12 from django.http import HttpResponse
  13
  14 from datetime import date
  15
  16 from django.core.urlresolvers import reverse
  17 from django.db import IntegrityError
  18
  19 import librarian
  20 import librarian.html
  21 import difflib
  22 from librarian import dcparser, parser
  23
  24 from wlrepo import *
  25 from api.models import PullRequest
  26 from explorer.models import GalleryForDocument
  27
  28 # internal imports
  29 import api.forms as forms
  30 import api.response as response
  31 from api.utils import validate_form, hglibrary, natural_order
  32 from api.models import PartCache, PullRequest
  33
  34 #
  35 import settings
  36
  37
  38 def is_prq(username):
  39     return username.startswith('$prq-')
  40
  41 def prq_for_user(username):
  42     try:
  43         return PullRequest.objects.get(id=int(username[5:]))
  44     except:
  45         return None
  46
  47 def check_user(request, user):
  48     log.info("user: %r, perm: %r" % (request.user, request.user.get_all_permissions()) )
  49     #pull request
  50     if is_prq(user):
  51         if not request.user.has_perm('api.view_prq'):
  52             yield response.AccessDenied().django_response({
  53                 'reason': 'access-denied',
  54                 'message': "You don't have enough priviliges to view pull requests."
  55             })
  56     # other users
  57     elif request.user.username != user:
  58         if not request.user.has_perm('api.view_other_document'):
  59             yield response.AccessDenied().django_response({
  60                 'reason': 'access-denied',
  61                 'message': "You don't have enough priviliges to view other people's document."
  62             })
  63     pass
  64
  65 #
  66 # Document List Handlers
  67 #
  68 # TODO: security check
  69 class BasicLibraryHandler(AnonymousBaseHandler):
  70     allowed_methods = ('GET',)
  71
  72     @hglibrary
  73     def read(self, request, lib):
  74         """Return the list of documents."""
  75         document_list = [{
  76             'url': reverse('document_view', args=[docid]),
  77             'name': docid } for docid in lib.documents() ]
  78         return {'documents' : document_list}
  79
  80 #
  81 # This handler controlls the document collection
  82 #
  83 class LibraryHandler(BaseHandler):
  84     allowed_methods = ('GET', 'POST')
  85     anonymous = BasicLibraryHandler
  86
  87     @hglibrary
  88     def read(self, request, lib):
  89         """Return the list of documents."""
  90
  91         documents = {}
  92
  93         for docid in lib.documents():
  94             documents[docid] = {
  95                 'url': reverse('document_view', args=[docid]),
  96                 'name': docid,
  97                 'parts': []
  98             }
  99
 100         parts = PartCache.objects.defer('part_id')\
 101             .values_list('part_id', 'document_id').distinct()
 102
 103         document_tree = dict(documents)
 104
 105         for part, docid in parts:
 106             # this way, we won't display broken links
 107             if not documents.has_key(part):
 108                 log.info("NOT FOUND: %s", part)
 109                 continue
 110
 111             parent = documents[docid]
 112             child = documents[part]
 113
 114             # not top-level anymore
 115             document_tree.pop(part)
 116             parent['parts'].append(child)
 117
 118         for doc in documents.itervalues():
 119             doc['parts'].sort(key=natural_order(lambda d: d['name']))
 120
 121         return {'documents': sorted(document_tree.itervalues(),
 122             key=natural_order(lambda d: d['name']) ) }
 123
 124
 125     @validate_form(forms.DocumentUploadForm, 'POST')
 126     @hglibrary
 127     def create(self, request, form, lib):
 128         """Create a new document."""
 129
 130         if form.cleaned_data['ocr_data']:
 131             data = form.cleaned_data['ocr_data']
 132         else:
 133             data = request.FILES['ocr_file'].read().decode('utf-8')
 134
 135         if data is None:
 136             return response.BadRequest().django_response('You must pass ocr_data or ocr_file.')
 137
 138         if form.cleaned_data['generate_dc']:
 139             data = librarian.wrap_text(data, unicode(date.today()))
 140
 141         docid = form.cleaned_data['bookname']
 142
 143         try:
 144             lock = lib.lock()
 145             try:
 146                 log.info("DOCID %s", docid)
 147                 doc = lib.document_create(docid)
 148                 # document created, but no content yet
 149                 try:
 150                     doc = doc.quickwrite('xml', data.encode('utf-8'),
 151                         '$AUTO$ XML data uploaded.', user=request.user.username)
 152                 except Exception,e:
 153                     import traceback
 154                     # rollback branch creation
 155                     lib._rollback()
 156                     raise LibraryException(traceback.format_exc())
 157
 158                 url = reverse('document_view', args=[doc.id])
 159
 160                 return response.EntityCreated().django_response(\
 161                     body = {
 162                         'url': url,
 163                         'name': doc.id,
 164                         'revision': doc.revision },
 165                     url = url )
 166             finally:
 167                 lock.release()
 168         except LibraryException, e:
 169             import traceback
 170             return response.InternalError().django_response({
 171                 "reason": traceback.format_exc()
 172             })
 173         except DocumentAlreadyExists:
 174             # Document is already there
 175             return response.EntityConflict().django_response({
 176                 "reason": "already-exists",
 177                 "message": "Document already exists." % docid
 178             })
 179
 180 #
 181 # Document Handlers
 182 #
 183 class BasicDocumentHandler(AnonymousBaseHandler):
 184     allowed_methods = ('GET',)
 185
 186     @hglibrary
 187     def read(self, request, docid, lib):
 188         try:
 189             doc = lib.document(docid)
 190         except RevisionNotFound:
 191             return rc.NOT_FOUND
 192
 193         result = {
 194             'name': doc.id,
 195             'html_url': reverse('dochtml_view', args=[doc.id]),
 196             'text_url': reverse('doctext_view', args=[doc.id]),
 197             'dc_url': reverse('docdc_view', args=[doc.id]),
 198             'public_revision': doc.revision,
 199         }
 200
 201         return result
 202
 203
 204 class DiffHandler(BaseHandler):
 205     allowed_methods = ('GET',)
 206
 207     @hglibrary
 208     def read(self, request, source_revision, target_revision, lib):
 209         '''Return diff between source_revision and target_revision)'''
 210         source_document = lib.document_for_rev(source_revision)
 211         target_document = lib.document_for_rev(target_revision)
 212         print source_document,
 213         print target_document
 214         diff = difflib.unified_diff(
 215             source_document.data('xml').splitlines(True),
 216             target_document.data('xml').splitlines(True),
 217             'source',
 218             'target')
 219
 220         return ''.join(list(diff))
 221
 222
 223 #
 224 # Document Meta Data
 225 #
 226 class DocumentHandler(BaseHandler):
 227     allowed_methods = ('GET', 'PUT')
 228     anonymous = BasicDocumentHandler
 229
 230     @validate_form(forms.DocumentRetrieveForm, 'GET')
 231     @hglibrary
 232     def read(self, request, form, docid, lib):
 233         """Read document's meta data"""
 234         log.info(u"User '%s' wants to %s(%s) as %s" % \
 235             (request.user.username, docid, form.cleaned_data['revision'], form.cleaned_data['user']) )
 236
 237         user = form.cleaned_data['user'] or request.user.username
 238         rev = form.cleaned_data['revision'] or 'latest'
 239
 240         for error in check_user(request, user):
 241             return error
 242
 243         try:
 244             doc = lib.document(docid, user, rev=rev)
 245         except RevisionMismatch, e:
 246             # the document exists, but the revision is bad
 247             return response.EntityNotFound().django_response({
 248                 'reason': 'revision-mismatch',
 249                 'message': e.message,
 250                 'docid': docid,
 251                 'user': user,
 252             })
 253         except RevisionNotFound, e:
 254             # the user doesn't have this document checked out
 255             # or some other weird error occured
 256             # try to do the checkout
 257             try:
 258                 if user == request.user.username:
 259                     mdoc = lib.document(docid)
 260                     doc = mdoc.take(user)
 261                 elif is_prq(user):
 262                     prq = prq_for_user(user)
 263                     # commiter's document
 264                     prq_doc = lib.document_for_rev(prq.source_revision)
 265                     doc = prq_doc.take(user)
 266                 else:
 267                     return response.EntityNotFound().django_response({
 268                         'reason': 'document-not-found',
 269                         'message': e.message,
 270                         'docid': docid,
 271                         'user': user,
 272                     })
 273             except RevisionNotFound, e:
 274                 return response.EntityNotFound().django_response({
 275                     'reason': 'document-not-found',
 276                     'message': e.message,
 277                     'docid': docid,
 278                     'user': user
 279                 })
 280
 281         return {
 282             'name': doc.id,
 283             'user': user,
 284             'html_url': reverse('dochtml_view', args=[doc.id]),
 285             'text_url': reverse('doctext_view', args=[doc.id]),
 286             # 'dc_url': reverse('docdc_view', args=[doc.id]),
 287             'gallery_url': reverse('docgallery_view', args=[doc.id]),
 288             'merge_url': reverse('docmerge_view', args=[doc.id]),
 289             'revision': doc.revision,
 290             'timestamp': doc.revision.timestamp,
 291             # 'public_revision': doc.revision,
 292             # 'public_timestamp': doc.revision.timestamp,
 293         }
 294
 295
 296 #    @hglibrary
 297 #    def update(self, request, docid, lib):
 298 #        """Update information about the document, like display not"""
 299 #        return
 300 #
 301 #
 302 #
 303 class DocumentHTMLHandler(BaseHandler):
 304     allowed_methods = ('GET')
 305
 306     @validate_form(forms.DocumentRetrieveForm, 'GET')
 307     @hglibrary
 308     def read(self, request, form, docid, lib, stylesheet='partial'):
 309         """Read document as html text"""
 310         try:
 311             revision = form.cleaned_data['revision']
 312             user = form.cleaned_data['user'] or request.user.username
 313             document = lib.document_for_rev(revision)
 314
 315             if document.id != docid:
 316                 return response.BadRequest().django_response({
 317                     'reason': 'name-mismatch',
 318                     'message': 'Provided revision is not valid for this document'
 319                 })
 320
 321             if document.owner != user:
 322                 return response.BadRequest().django_response({
 323                     'reason': 'user-mismatch',
 324                     'message': "Provided revision doesn't belong to user %s" % user
 325                 })
 326
 327             for error in check_user(request, user):
 328                 return error
 329
 330             return librarian.html.transform(document.data('xml'), is_file=False, \
 331                 parse_dublincore=False, stylesheet=stylesheet,\
 332                 options={
 333                     "with-paths": 'boolean(1)',
 334                 })
 335
 336         except (EntryNotFound, RevisionNotFound), e:
 337             return response.EntityNotFound().django_response({
 338                 'reason': 'not-found', 'message': e.message})
 339         except librarian.ParseError, e:
 340             return response.InternalError().django_response({
 341                 'reason': 'xml-parse-error', 'message': e.message })
 342
 343 #
 344 # Image Gallery
 345 #
 346
 347 class DocumentGalleryHandler(BaseHandler):
 348     allowed_methods = ('GET')
 349
 350
 351     def read(self, request, docid):
 352         """Read meta-data about scans for gallery of this document."""
 353         galleries = []
 354         from urllib import quote
 355
 356         for assoc in GalleryForDocument.objects.filter(document=docid):
 357             dirpath = os.path.join(settings.MEDIA_ROOT, assoc.subpath)
 358
 359             if not os.path.isdir(dirpath):
 360                 log.warn(u"[WARNING]: missing gallery %s", dirpath)
 361                 continue
 362
 363             gallery = {'name': assoc.name, 'pages': []}
 364
 365             for file in os.listdir(dirpath):
 366                 if not isinstance(file, unicode):
 367                     try:
 368                         file = file.decode('utf-8')
 369                     except:
 370                         log.warn(u"File %r in gallery %r is not unicode. Ommiting."\
 371                             % (file, dirpath) )
 372                         file = None
 373
 374                 if file is not None:
 375                     name, ext = os.path.splitext(os.path.basename(file))
 376
 377                     if ext.lower() not in [u'.png', u'.jpeg', u'.jpg']:
 378                         log.warn(u"Ignoring: %s %s", name, ext)
 379                         url = None
 380
 381                     url = settings.MEDIA_URL + assoc.subpath + u'/' + file
 382
 383                 if url is None:
 384                     url = settings.MEDIA_URL + u'/missing.png'
 385
 386                 gallery['pages'].append( quote(url.encode('utf-8')) )
 387
 388             gallery['pages'].sort()
 389             galleries.append(gallery)
 390
 391         return galleries
 392
 393
 394
 395 #
 396 # Dublin Core handlers
 397 #
 398 # @requires librarian
 399 #
 400 #class DocumentDublinCoreHandler(BaseHandler):
 401 #    allowed_methods = ('GET', 'POST')
 402 #
 403 #    @hglibrary
 404 #    def read(self, request, docid, lib):
 405 #        """Read document as raw text"""
 406 #        try:
 407 #            revision = request.GET.get('revision', 'latest')
 408 #
 409 #            if revision == 'latest':
 410 #                doc = lib.document(docid)
 411 #            else:
 412 #                doc = lib.document_for_rev(revision)
 413 #
 414 #
 415 #            if document.id != docid:
 416 #                return response.BadRequest().django_response({'reason': 'name-mismatch',
 417 #                    'message': 'Provided revision is not valid for this document'})
 418 #
 419 #            bookinfo = dcparser.BookInfo.from_string(doc.data('xml'))
 420 #            return bookinfo.serialize()
 421 #        except (EntryNotFound, RevisionNotFound), e:
 422 #            return response.EntityNotFound().django_response({
 423 #                'exception': type(e), 'message': e.message})
 424 #
 425 #    @hglibrary
 426 #    def create(self, request, docid, lib):
 427 #        try:
 428 #            bi_json = request.POST['contents']
 429 #            revision = request.POST['revision']
 430 #
 431 #            if request.POST.has_key('message'):
 432 #                msg = u"$USER$ " + request.PUT['message']
 433 #            else:
 434 #                msg = u"$AUTO$ Dublin core update."
 435 #
 436 #            current = lib.document(docid, request.user.username)
 437 #            orig = lib.document_for_rev(revision)
 438 #
 439 #            if current != orig:
 440 #                return response.EntityConflict().django_response({
 441 #                        "reason": "out-of-date",
 442 #                        "provided": orig.revision,
 443 #                        "latest": current.revision })
 444 #
 445 #            xmldoc = parser.WLDocument.from_string(current.data('xml'))
 446 #            document.book_info = dcparser.BookInfo.from_json(bi_json)
 447 #
 448 #            # zapisz
 449 #            ndoc = current.quickwrite('xml', \
 450 #                document.serialize().encode('utf-8'),\
 451 #                message=msg, user=request.user.username)
 452 #
 453 #            try:
 454 #                # return the new revision number
 455 #                return {
 456 #                    "document": ndoc.id,
 457 #                    "subview": "dc",
 458 #                    "previous_revision": current.revision,
 459 #                    "revision": ndoc.revision,
 460 #                    'timestamp': ndoc.revision.timestamp,
 461 #                    "url": reverse("docdc_view", args=[ndoc.id])
 462 #                }
 463 #            except Exception, e:
 464 #                if ndoc: lib._rollback()
 465 #                raise e
 466 #        except RevisionNotFound:
 467 #            return response.EntityNotFound().django_response()
 468
 469 class MergeHandler(BaseHandler):
 470     allowed_methods = ('POST',)
 471
 472     @validate_form(forms.MergeRequestForm, 'POST')
 473     @hglibrary
 474     def create(self, request, form, docid, lib):
 475         """Create a new document revision from the information provided by user"""
 476         revision = form.cleaned_data['revision']
 477
 478         # fetch the main branch document
 479         doc = lib.document(docid)
 480
 481         # fetch the base document
 482         user_doc = lib.document_for_rev(revision)
 483         base_doc = user_doc.latest()
 484
 485         if base_doc != user_doc:
 486             return response.EntityConflict().django_response({
 487                 "reason": "out-of-date",
 488                 "provided": str(user_doc.revision),
 489                 "latest": str(base_doc.revision)
 490             })
 491
 492         if form.cleaned_data['type'] == 'update':
 493             # update is always performed from the file branch
 494             # to the user branch
 495             user_doc_new = base_doc.update(request.user.username)
 496
 497             if user_doc_new == user_doc:
 498                 return response.SuccessAllOk().django_response({
 499                     "result": "no-op"
 500                 })
 501
 502             # shared document is the same
 503             doc_new = doc
 504
 505         if form.cleaned_data['type'] == 'share':
 506             if not base_doc.up_to_date():
 507                 return response.BadRequest().django_response({
 508                     "reason": "not-fast-forward",
 509                     "message": "You must first update your branch to the latest version."
 510                 })
 511
 512             if base_doc.parentof(doc) or base_doc.has_parent_from(doc):
 513                 return response.SuccessAllOk().django_response({
 514                     "result": "no-op"
 515                 })
 516
 517             # check for unresolved conflicts
 518             if base_doc.has_conflict_marks():
 519                 return response.BadRequest().django_response({
 520                     "reason": "unresolved-conflicts",
 521                     "message": "There are unresolved conflicts in your file. Fix them, and try again."
 522                 })
 523
 524             if not request.user.has_perm('api.share_document'):
 525                 # User is not permitted to make a merge, right away
 526                 # So we instead create a pull request in the database
 527                 try:
 528                     prq, created = PullRequest.objects.get_or_create(
 529                         comitter = request.user,
 530                         document = docid,
 531                         status = "N",
 532                         defaults = {
 533                             'source_revision': str(base_doc.revision),
 534                             'comment': form.cleaned_data['message'] or '$AUTO$ Document shared.',
 535                         }
 536                     )
 537
 538                     # there can't be 2 pending request from same user
 539                     # for the same document
 540                     if not created:
 541                         prq.source_revision = str(base_doc.revision)
 542                         prq.comment = prq.comment + 'u\n\n' + (form.cleaned_data['message'] or u'')
 543                         prq.save()
 544
 545                     return response.RequestAccepted().django_response(\
 546                         ticket_status=prq.status, \
 547                         ticket_uri=reverse("pullrequest_view", args=[prq.id]) )
 548                 except IntegrityError:
 549                     return response.EntityConflict().django_response({
 550                         'reason': 'request-already-exist'
 551                     })
 552
 553             changed = base_doc.share(form.cleaned_data['message'])
 554
 555             # update shared version if needed
 556             if changed:
 557                 doc_new = doc.latest()
 558             else:
 559                 doc_new = doc
 560
 561             # the user wersion is the same
 562             user_doc_new = base_doc
 563
 564         # The client can compare parent_revision to revision
 565         # to see if he needs to update user's view
 566         # Same goes for shared view
 567
 568         return response.SuccessAllOk().django_response({
 569             "result": "success",
 570             "name": user_doc_new.id,
 571             "user": user_doc_new.owner,
 572
 573             "revision": user_doc_new.revision,
 574             'timestamp': user_doc_new.revision.timestamp,
 575
 576             "parent_revision": user_doc.revision,
 577             "parent_timestamp": user_doc.revision.timestamp,
 578
 579             "shared_revision": doc_new.revision,
 580             "shared_timestamp": doc_new.revision.timestamp,
 581
 582             "shared_parent_revision": doc.revision,
 583             "shared_parent_timestamp": doc.revision.timestamp,
 584         })