Child: [bea16d] (diff)

Download this file

teamforge-import.py    1035 lines (925 with data), 44.4 kB

   1
   2
   3
   4
   5
   6
   7
   8
   9
  10
  11
  12
  13
  14
  15
  16
  17
  18
  19
  20
  21
  22
  23
  24
  25
  26
  27
  28
  29
  30
  31
  32
  33
  34
  35
  36
  37
  38
  39
  40
  41
  42
  43
  44
  45
  46
  47
  48
  49
  50
  51
  52
  53
  54
  55
  56
  57
  58
  59
  60
  61
  62
  63
  64
  65
  66
  67
  68
  69
  70
  71
  72
  73
  74
  75
  76
  77
  78
  79
  80
  81
  82
  83
  84
  85
  86
  87
  88
  89
  90
  91
  92
  93
  94
  95
  96
  97
  98
  99
 100
 101
 102
 103
 104
 105
 106
 107
 108
 109
 110
 111
 112
 113
 114
 115
 116
 117
 118
 119
 120
 121
 122
 123
 124
 125
 126
 127
 128
 129
 130
 131
 132
 133
 134
 135
 136
 137
 138
 139
 140
 141
 142
 143
 144
 145
 146
 147
 148
 149
 150
 151
 152
 153
 154
 155
 156
 157
 158
 159
 160
 161
 162
 163
 164
 165
 166
 167
 168
 169
 170
 171
 172
 173
 174
 175
 176
 177
 178
 179
 180
 181
 182
 183
 184
 185
 186
 187
 188
 189
 190
 191
 192
 193
 194
 195
 196
 197
 198
 199
 200
 201
 202
 203
 204
 205
 206
 207
 208
 209
 210
 211
 212
 213
 214
 215
 216
 217
 218
 219
 220
 221
 222
 223
 224
 225
 226
 227
 228
 229
 230
 231
 232
 233
 234
 235
 236
 237
 238
 239
 240
 241
 242
 243
 244
 245
 246
 247
 248
 249
 250
 251
 252
 253
 254
 255
 256
 257
 258
 259
 260
 261
 262
 263
 264
 265
 266
 267
 268
 269
 270
 271
 272
 273
 274
 275
 276
 277
 278
 279
 280
 281
 282
 283
 284
 285
 286
 287
 288
 289
 290
 291
 292
 293
 294
 295
 296
 297
 298
 299
 300
 301
 302
 303
 304
 305
 306
 307
 308
 309
 310
 311
 312
 313
 314
 315
 316
 317
 318
 319
 320
 321
 322
 323
 324
 325
 326
 327
 328
 329
 330
 331
 332
 333
 334
 335
 336
 337
 338
 339
 340
 341
 342
 343
 344
 345
 346
 347
 348
 349
 350
 351
 352
 353
 354
 355
 356
 357
 358
 359
 360
 361
 362
 363
 364
 365
 366
 367
 368
 369
 370
 371
 372
 373
 374
 375
 376
 377
 378
 379
 380
 381
 382
 383
 384
 385
 386
 387
 388
 389
 390
 391
 392
 393
 394
 395
 396
 397
 398
 399
 400
 401
 402
 403
 404
 405
 406
 407
 408
 409
 410
 411
 412
 413
 414
 415
 416
 417
 418
 419
 420
 421
 422
 423
 424
 425
 426
 427
 428
 429
 430
 431
 432
 433
 434
 435
 436
 437
 438
 439
 440
 441
 442
 443
 444
 445
 446
 447
 448
 449
 450
 451
 452
 453
 454
 455
 456
 457
 458
 459
 460
 461
 462
 463
 464
 465
 466
 467
 468
 469
 470
 471
 472
 473
 474
 475
 476
 477
 478
 479
 480
 481
 482
 483
 484
 485
 486
 487
 488
 489
 490
 491
 492
 493
 494
 495
 496
 497
 498
 499
 500
 501
 502
 503
 504
 505
 506
 507
 508
 509
 510
 511
 512
 513
 514
 515
 516
 517
 518
 519
 520
 521
 522
 523
 524
 525
 526
 527
 528
 529
 530
 531
 532
 533
 534
 535
 536
 537
 538
 539
 540
 541
 542
 543
 544
 545
 546
 547
 548
 549
 550
 551
 552
 553
 554
 555
 556
 557
 558
 559
 560
 561
 562
 563
 564
 565
 566
 567
 568
 569
 570
 571
 572
 573
 574
 575
 576
 577
 578
 579
 580
 581
 582
 583
 584
 585
 586
 587
 588
 589
 590
 591
 592
 593
 594
 595
 596
 597
 598
 599
 600
 601
 602
 603
 604
 605
 606
 607
 608
 609
 610
 611
 612
 613
 614
 615
 616
 617
 618
 619
 620
 621
 622
 623
 624
 625
 626
 627
 628
 629
 630
 631
 632
 633
 634
 635
 636
 637
 638
 639
 640
 641
 642
 643
 644
 645
 646
 647
 648
 649
 650
 651
 652
 653
 654
 655
 656
 657
 658
 659
 660
 661
 662
 663
 664
 665
 666
 667
 668
 669
 670
 671
 672
 673
 674
 675
 676
 677
 678
 679
 680
 681
 682
 683
 684
 685
 686
 687
 688
 689
 690
 691
 692
 693
 694
 695
 696
 697
 698
 699
 700
 701
 702
 703
 704
 705
 706
 707
 708
 709
 710
 711
 712
 713
 714
 715
 716
 717
 718
 719
 720
 721
 722
 723
 724
 725
 726
 727
 728
 729
 730
 731
 732
 733
 734
 735
 736
 737
 738
 739
 740
 741
 742
 743
 744
 745
 746
 747
 748
 749
 750
 751
 752
 753
 754
 755
 756
 757
 758
 759
 760
 761
 762
 763
 764
 765
 766
 767
 768
 769
 770
 771
 772
 773
 774
 775
 776
 777
 778
 779
 780
 781
 782
 783
 784
 785
 786
 787
 788
 789
 790
 791
 792
 793
 794
 795
 796
 797
 798
 799
 800
 801
 802
 803
 804
 805
 806
 807
 808
 809
 810
 811
 812
 813
 814
 815
 816
 817
 818
 819
 820
 821
 822
 823
 824
 825
 826
 827
 828
 829
 830
 831
 832
 833
 834
 835
 836
 837
 838
 839
 840
 841
 842
 843
 844
 845
 846
 847
 848
 849
 850
 851
 852
 853
 854
 855
 856
 857
 858
 859
 860
 861
 862
 863
 864
 865
 866
 867
 868
 869
 870
 871
 872
 873
 874
 875
 876
 877
 878
 879
 880
 881
 882
 883
 884
 885
 886
 887
 888
 889
 890
 891
 892
 893
 894
 895
 896
 897
 898
 899
 900
 901
 902
 903
 904
 905
 906
 907
 908
 909
 910
 911
 912
 913
 914
 915
 916
 917
 918
 919
 920
 921
 922
 923
 924
 925
 926
 927
 928
 929
 930
 931
 932
 933
 934
 935
 936
 937
 938
 939
 940
 941
 942
 943
 944
 945
 946
 947
 948
 949
 950
 951
 952
 953
 954
 955
 956
 957
 958
 959
 960
 961
 962
 963
 964
 965
 966
 967
 968
 969
 970
 971
 972
 973
 974
 975
 976
 977
 978
 979
 980
 981
 982
 983
 984
 985
 986
 987
 988
 989
 990
 991
 992
 993
 994
 995
 996
 997
 998
 999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
import logging
from getpass import getpass
from optparse import OptionParser
from pylons import c
import re
import os
from time import mktime
import time
import json
from urlparse import urlparse
import urllib2
import urllib
from cookielib import CookieJar
from datetime import datetime
from ConfigParser import ConfigParser
import random
import string
import sqlalchemy
from suds.client import Client
from ming.orm.ormsession import ThreadLocalORMSession
from ming.base import Object
from allura import model as M
from allura.lib import helpers as h
from allura.lib import utils
log = logging.getLogger('teamforge-import')
'''
http://help.collab.net/index.jsp?topic=/teamforge520/reference/api-services.html
http://www.open.collab.net/nonav/community/cif/csfe/50/javadoc/index.html?com/collabnet/ce/soap50/webservices/page/package-summary.html
'''
options = None
s = None # security token
client = None # main api client
users = {}
cj = CookieJar()
loggedInOpener = urllib2.build_opener(urllib2.HTTPCookieProcessor(cj))
def make_client(api_url, app):
return Client(api_url + app + '?wsdl', location=api_url + app)
def main():
global options, s, client, users
defaults=dict(
api_url=None,
attachment_url='/sf/%s/do/%s/',
default_wiki_text='PRODUCT NAME HERE',
username=None,
password=None,
output_dir='teamforge-export/',
list_project_ids=False,
neighborhood=None,
neighborhood_shortname=None,
use_thread_import_id_when_reloading=False,
skip_wiki=False,
skip_frs_download=False,
skip_unsupported_check=False)
optparser = get_parser(defaults)
options, project_ids = optparser.parse_args()
if options.config_file:
config = ConfigParser()
config.read(options.config_file)
defaults.update(
(k, eval(v)) for k,v in config.items('teamforge-import'))
optparser = get_parser(defaults)
options, project_ids = optparser.parse_args()
# neither specified, so do both
if not options.extract and not options.load:
options.extract = True
options.load = True
if options.extract:
client = make_client(options.api_url, 'CollabNet')
api_v = client.service.getApiVersion()
if not api_v.startswith('5.4.'):
log.warning('Unexpected API Version %s. May not work correctly.' % api_v)
s = client.service.login(options.username, options.password or getpass('Password: '))
teamforge_v = client.service.getVersion(s)
if not teamforge_v.startswith('5.4.'):
log.warning('Unexpected TeamForge Version %s. May not work correctly.' % teamforge_v)
if options.load:
if not options.neighborhood:
log.error('You must specify a neighborhood when loading')
return
try:
nbhd = M.Neighborhood.query.get(name=options.neighborhood)
except:
log.exception('error querying mongo')
log.error('This should be run as "paster script production.ini ../scripts/teamforge-import.py -- ...options.."')
return
assert nbhd
if not project_ids:
if not options.extract:
log.error('You must specify project ids')
return
projects = client.service.getProjectList(s)
project_ids = [p.id for p in projects.dataRows]
if options.list_project_ids:
print ' '.join(project_ids)
return
if not os.path.exists(options.output_dir):
os.makedirs(options.output_dir)
for pid in project_ids:
if options.extract:
try:
project = client.service.getProjectData(s, pid)
log.info('Project: %s %s %s' % (project.id, project.title, project.path))
out_dir = os.path.join(options.output_dir, project.id)
if not os.path.exists(out_dir):
os.mkdir(out_dir)
get_project(project)
get_files(project)
if not options.skip_wiki:
get_homepage_wiki(project)
get_discussion(project)
get_news(project)
if not options.skip_unsupported_check:
check_unsupported_tools(project)
with open(os.path.join(options.output_dir, 'users.json'), 'w') as user_file:
json.dump(users, user_file, default=str)
except:
log.exception('Error extracting %s' % pid)
if options.load:
try:
project = create_project(pid, nbhd)
except:
log.exception('Error creating %s' % pid)
def load_users():
''' load the users data from file, if it hasn't been already '''
global users
user_filename = os.path.join(options.output_dir, 'users.json')
if not users and os.path.exists(user_filename):
with open(user_filename) as user_file:
users = json.load(user_file, object_hook=Object) # Object for attribute access
def save_user(usernames):
if isinstance(usernames, basestring):
usernames = [usernames]
load_users()
for username in usernames:
if username not in users:
user_data = client.service.getUserData(s, username)
users[username] = Object(user_data)
if users[username].status != 'Active':
log.warn('user: %s status: %s' % (username, users[username].status))
def get_project(project):
global client
cats = make_client(options.api_url, 'CategorizationApp')
data = client.service.getProjectData(s, project.id)
access_level = { 1: 'public', 4: 'private', 3: 'gated community'}[
client.service.getProjectAccessLevel(s, project.id)
]
admins = client.service.listProjectAdmins(s, project.id).dataRows
members = client.service.getProjectMemberList(s, project.id).dataRows
groups = client.service.getProjectGroupList(s, project.id).dataRows
categories = cats.service.getProjectCategories(s, project.id).dataRows
save(json.dumps(dict(
data = dict(data),
access_level = access_level,
admins = map(dict, admins),
members = map(dict, members),
groups = map(dict, groups),
categories = map(dict, categories),
), default=str),
project, project.id+'.json')
if len(groups):
log.warn('Project has groups %s' % groups)
for u in admins:
if not u.status != 'active':
log.warn('inactive admin %s' % u)
if u.superUser:
log.warn('super user admin %s' % u)
save_user(data.createdBy)
save_user(u.userName for u in admins)
save_user(u.userName for u in members)
def get_user(orig_username):
'returns an allura User object'
sf_username = make_valid_sf_username(orig_username)
u = M.User.by_username(sf_username)
if not u:
load_users()
user = users[orig_username]
if user.status != 'Active':
log.warn('Inactive user %s %s' % (orig_username, user.status))
if not 3 <= len(user.fullName) <= 32:
raise Exception('invalid fullName length: %s' % user.fullName)
if '@' not in user.email:
raise Exception('invalid email: %s' % user.email)
# FIXME: hardcoded SFX integration
from sfx.model import tables as T
nu = T.users.insert()
nu.execute(user_name=sf_username.encode('utf-8'),
email=user.email.lower().encode('utf-8'),
realname=user.fullName.encode('utf-8'),
status='A' if user.status == 'Active' else 'D',
language=275, # english trove id
timezone=user.timeZone,
user_pw=''.join(random.sample(string.printable, 32)),
unix_pw=''.join(random.sample(string.printable, 32)),
user_pw_modtime=int(time.time()),
mail_siteupdates=0,
add_date=int(time.time()),
)
user_id = sqlalchemy.select([T.users.c.user_id], T.users.c.user_name==sf_username).execute().fetchone().user_id
npref = T.user_preferences.insert()
npref.execute(user_id=user_id, preference_name='country', preference_value='US')
npref.execute(user_id=user_id, preference_name='opt_research', preference_value=0)
npref.execute(user_id=user_id, preference_name='opt_thirdparty', preference_value=0)
new_audit = T.audit_trail_user.insert()
new_audit.execute(
date=int(time.time()),
username='nobody',
ip_address='(imported)',
operation_resource=user_id,
operation='%s user account created by TeamForge import script' % user.status,
operation_target='',
)
u = M.User.by_username(sf_username)
assert u
return u
def convert_project_shortname(teamforge_path):
'convert from TeamForge to SF, and validate early'
tf_shortname = teamforge_path.split('.')[-1]
sf_shortname = tf_shortname.replace('_','-')
# FIXME hardcoded translations
sf_shortname = {
'i1': 'motorola-i1',
'i9': 'motorola-i9',
'devplatformforocap': 'ocap-dev-pltfrm',
'sitewide': '--init--',
}.get(sf_shortname, sf_shortname)
if not 3 <= len(sf_shortname) <= 15:
raise ValueError('Project name length must be between 3 & 15, inclusive: %s (%s)' %
(sf_shortname, len(sf_shortname)))
return sf_shortname
# FIXME hardcoded
skip_perms_usernames = set([
'faisal_saeed','dsarkisian','debonairamit','nishanthiremath','Bhuvnesh','bluetooth','cnkurzke','makow2','jannes1','Joel_Hegberg','Farroc','brian_chen','eirikur',
'dmitry_flyorov','bipingm','MornayJo','ibv','b_weisshaar','k9srb','johnmmills','a_gomolitsky','filim','kapoor','ljzegers','jrukes','dwilson9','jlin','quickie',
'johnbell','nnikolenko','Gaetan','Giannetta','Katia','jackhan','jacobwangus','adwankar','dinobrusco','qbarnes','ilmojung','clifford_chan','nbaig','fhutchi1',
'rinofarina','baiyanbin','muralidhar','duanyiruo','bredding','mkolkey','manvith','nanduk','engyihan','deepsie','dabon','dino_jiang','mattrose','peter_j_wilhelm',
'emx2500','jmcguire','lfilimowski','guruppandit','abhilashisme','edwinhm','rabbi','ferrans','guna','kevin_robinson','adathiruthi','kochen','onehap','kalanithi',
'jamesn','obu001','chetanv','Avinash','HugoBoss','Han_Wei','mhooper','g16872','mfcarignano','jim_burke','kevin','arunkarra','adam_feng','pavan_scm','kostya_katz',
'ppazderka','eileenzhuang','pyammine','judyho','ashoykh','rdemento','ibrahim','min_wang','arvind_setlur','moorthy_karthik','daniel_nelson','dms','esnmurthy',
'rasa_bonyadlou','prashantjoshi','edkeating','billsaez','cambalindo','jims','bozkoyun','andry_deltsov','bpowers','manuel_milli','maryparsons','spriporov','yutianli',
'xiebin','tnemeth1','udayaps','zzzzuser','timberger','sbarve1','zarman','rwallace67','thangavelu_arum','yuhuaixie','tingup','sekchai','sasanplus','rupal','sebastien_hertz',
'sab8123','rony_lim','slava_kirillin','smwest','wendydu_yq','sco002','RonFred','spatnala','vd','Sunny','tthompson','sunijams','slaw','rodovich','zhangqingqi82','venki',
'yuntaom','xiaojin','walterciocosta','straus','Thomas','stupka','wangyu','yaowang','wisekb','tyler_louie','smartgarfield','shekar_mahalingam',
'venkata_akella','v_yellapragada','vavasthi','rpatel','zhengfang','sweetybala','vap','sergey','ymhuang','spatel78745'
])
def create_project(pid, nbhd):
M.session.artifact_orm_session._get().skip_mod_date = True
data = loadjson(pid, pid+'.json')
#pprint(data)
log.info('Loading: %s %s %s' % (pid, data.data.title, data.data.path))
shortname = convert_project_shortname(data.data.path)
project = M.Project.query.get(shortname=shortname, neighborhood_id=nbhd._id)
if not project:
private = (data.access_level == 'private')
log.debug('Creating %s private=%s' % (shortname, private))
one_admin = [u.userName for u in data.admins if u.status == 'Active'][0]
project = nbhd.register_project(shortname,
get_user(one_admin),
project_name=data.data.title,
private_project=private)
project.notifications_disabled = True
project.short_description = data.data.description
project.last_updated = datetime.strptime(data.data.lastModifiedDate, '%Y-%m-%d %H:%M:%S')
M.main_orm_session.flush(project)
# TODO: push last_updated to gutenberg?
# TODO: try to set createdDate?
role_admin = M.ProjectRole.by_name('Admin', project)
admin_usernames = set()
for admin in data.admins:
# FIXME: skip non-active users
if admin.userName in skip_perms_usernames:
continue
admin_usernames.add(admin.userName)
user = get_user(admin.userName)
c.user = user
pr = user.project_role(project)
pr.roles = [ role_admin._id ]
ThreadLocalORMSession.flush_all()
role_developer = M.ProjectRole.by_name('Developer', project)
for member in data.members:
# FIXME: skip non-active users
if member.userName in skip_perms_usernames:
continue
if member.userName in admin_usernames:
continue
user = get_user(member.userName)
pr = user.project_role(project)
pr.roles = [ role_developer._id ]
ThreadLocalORMSession.flush_all()
project.labels = [cat.path.split('projects/categorization.root.')[1] for cat in data.categories]
icon_file = 'emsignia-MOBILITY-red.png'
if 'nsn' in project.labels or 'msi' in project.labels:
icon_file = 'emsignia-SOLUTIONS-blue.gif'
if project.icon:
M.ProjectFile.remove(dict(project_id=project._id, category='icon'))
with open(os.path.join('..','scripts',icon_file)) as fp:
M.ProjectFile.save_image(
icon_file, fp, content_type=utils.guess_mime_type(icon_file),
square=True, thumbnail_size=(48,48),
thumbnail_meta=dict(project_id=project._id,category='icon'))
ThreadLocalORMSession.flush_all()
dirs = os.listdir(os.path.join(options.output_dir, pid))
frs_mapping = loadjson(pid, 'frs_mapping.json')
if not options.skip_wiki and 'wiki' in dirs:
import_wiki(project, pid, nbhd)
if not options.skip_frs_download and not project.app_instance('downloads'):
project.install_app('Downloads', 'downloads')
if 'forum' in dirs:
import_discussion(project, pid, frs_mapping, shortname, nbhd)
if 'news' in dirs:
import_news(project, pid, frs_mapping, shortname, nbhd)
project.notifications_disabled = False
ThreadLocalORMSession.flush_all()
return project
def import_wiki(project, pid, nbhd):
from forgewiki import model as WM
def upload_attachments(page, pid, beginning):
dirpath = os.path.join(options.output_dir, pid, 'wiki', beginning)
if not os.path.exists(dirpath): return
files = os.listdir(dirpath)
for f in files:
with open(os.path.join(options.output_dir, pid, 'wiki', beginning, f)) as fp:
page.attach(f, fp, content_type=utils.guess_mime_type(f))
pages = os.listdir(os.path.join(options.output_dir, pid, 'wiki'))
# handle the homepage content
if 'homepage_text.markdown' in pages:
home_app = project.app_instance('home')
h.set_context(project.shortname, 'home', neighborhood=nbhd)
# set permissions and config options
role_admin = M.ProjectRole.by_name('Admin')._id
role_anon = M.ProjectRole.by_name('*anonymous')._id
home_app.config.options['show_discussion'] = False
home_app.config.options['show_left_bar'] = False
home_app.config.options['show_right_bar'] = False
home_app.config.acl = [
M.ACE.allow(role_anon, 'read'),
M.ACE.allow(role_admin, 'create'),
M.ACE.allow(role_admin, 'edit'),
M.ACE.allow(role_admin, 'delete'),
M.ACE.allow(role_admin, 'moderate'),
M.ACE.allow(role_admin, 'configure'),
M.ACE.allow(role_admin, 'admin')]
p = WM.Page.upsert('Home')
p.text = wiki2markdown(load(pid, 'wiki', 'homepage_text.markdown'))
upload_attachments(p, pid, 'homepage')
if 'HomePage.json' in pages and 'HomePage.markdown' in pages:
wiki_app = project.app_instance('wiki')
if not wiki_app:
wiki_app = project.install_app('Wiki', 'wiki')
h.set_context(project.shortname, 'wiki', neighborhood=nbhd)
# set permissions and config options
role_admin = M.ProjectRole.by_name('Admin')._id
role_anon = M.ProjectRole.by_name('*anonymous')._id
wiki_app.config.options['show_discussion'] = False
wiki_app.config.options['show_left_bar'] = False
wiki_app.config.options['show_right_bar'] = False
wiki_app.config.acl = [
M.ACE.allow(role_anon, 'read'),
M.ACE.allow(role_admin, 'create'),
M.ACE.allow(role_admin, 'edit'),
M.ACE.allow(role_admin, 'delete'),
M.ACE.allow(role_admin, 'moderate'),
M.ACE.allow(role_admin, 'configure'),
M.ACE.allow(role_admin, 'admin')]
# make all the wiki pages
for page in pages:
ending = page[-5:]
beginning = page[:-5]
markdown_file = '%s.markdown' % beginning
if '.json' == ending and markdown_file in pages:
page_data = loadjson(pid, 'wiki', page)
content = load(pid, 'wiki', markdown_file)
if page == 'HomePage.json':
globals = WM.Globals.query.get(app_config_id=wiki_app.config._id)
if globals is not None:
globals.root = page_data.title
else:
globals = WM.Globals(app_config_id=wiki_app.config._id, root=page_data.title)
p = WM.Page.upsert(page_data.title)
p.viewable_by = ['all']
p.text = wiki2markdown(content)
# upload attachments
upload_attachments(p, pid, beginning)
if not p.history().first():
p.commit()
ThreadLocalORMSession.flush_all()
def import_discussion(project, pid, frs_mapping, sf_project_shortname, nbhd):
from forgediscussion import model as DM
discuss_app = project.app_instance('discussion')
if not discuss_app:
discuss_app = project.install_app('Discussion', 'discussion')
h.set_context(project.shortname, 'discussion', neighborhood=nbhd)
assert c.app
# set permissions and config options
role_admin = M.ProjectRole.by_name('Admin')._id
role_developer = M.ProjectRole.by_name('Developer')._id
role_auth = M.ProjectRole.by_name('*authenticated')._id
role_anon = M.ProjectRole.by_name('*anonymous')._id
discuss_app.config.acl = [
M.ACE.allow(role_anon, 'read'),
M.ACE.allow(role_auth, 'post'),
M.ACE.allow(role_auth, 'unmoderated_post'),
M.ACE.allow(role_developer, 'moderate'),
M.ACE.allow(role_admin, 'configure'),
M.ACE.allow(role_admin, 'admin')]
ThreadLocalORMSession.flush_all()
DM.Forum.query.remove(dict(app_config_id=discuss_app.config._id,shortname='general'))
forums = os.listdir(os.path.join(options.output_dir, pid, 'forum'))
for forum in forums:
ending = forum[-5:]
forum_name = forum[:-5]
if '.json' == ending and forum_name in forums:
forum_data = loadjson(pid, 'forum', forum)
fo = DM.Forum.query.get(shortname=forum_name, app_config_id=discuss_app.config._id)
if not fo:
fo = DM.Forum(app_config_id=discuss_app.config._id, shortname=forum_name)
fo.name = forum_data.title
fo.description = forum_data.description
fo_num_topics = 0
fo_num_posts = 0
topics = os.listdir(os.path.join(options.output_dir, pid, 'forum', forum_name))
for topic in topics:
ending = topic[-5:]
topic_name = topic[:-5]
if '.json' == ending and topic_name in topics:
fo_num_topics += 1
topic_data = loadjson(pid, 'forum', forum_name, topic)
thread_query = dict(
subject=topic_data.title,
discussion_id=fo._id,
app_config_id=discuss_app.config._id)
if not options.skip_thread_import_id_when_reloading:
# temporary/transitional. Just needed the first time
# running with this new code against an existing import
# that didn't have import_ids
thread_query['import_id'] = topic_data.id
to = DM.ForumThread.query.get(**thread_query)
if not to:
to = DM.ForumThread.new(
subject=topic_data.title,
discussion_id=fo._id,
import_id=topic_data.id,
app_config_id=discuss_app.config._id)
to.import_id=topic_data.id
to_num_replies = 0
oldest_post = None
newest_post = None
posts = sorted(os.listdir(os.path.join(options.output_dir, pid, 'forum', forum_name, topic_name)))
for post in posts:
ending = post[-5:]
post_name = post[:-5]
if '.json' == ending:
to_num_replies += 1
post_data = loadjson(pid, 'forum', forum_name, topic_name, post)
p = DM.ForumPost.query.get(
_id='%s%s@import' % (post_name,str(discuss_app.config._id)),
thread_id=to._id,
discussion_id=fo._id,
app_config_id=discuss_app.config._id)
if not p:
p = DM.ForumPost(
_id='%s%s@import' % (post_name,str(discuss_app.config._id)),
thread_id=to._id,
discussion_id=fo._id,
app_config_id=discuss_app.config._id)
create_date = datetime.strptime(post_data.createdDate, '%Y-%m-%d %H:%M:%S')
p.timestamp = create_date
p.author_id = str(get_user(post_data.createdByUserName)._id)
p.text = convert_post_content(frs_mapping, sf_project_shortname, post_data.content, nbhd)
p.status = 'ok'
if post_data.replyToId:
p.parent_id = '%s%s@import' % (post_data.replyToId,str(discuss_app.config._id))
slug, full_slug = p.make_slugs(parent = p.parent, timestamp = create_date)
p.slug = slug
p.full_slug = full_slug
if oldest_post == None or oldest_post.timestamp > create_date:
oldest_post = p
if newest_post == None or newest_post.timestamp < create_date:
newest_post = p
ThreadLocalORMSession.flush_all()
to.num_replies = to_num_replies
to.first_post_id = oldest_post._id
to.last_post_date = newest_post.timestamp
to.mod_date = newest_post.timestamp
fo_num_posts += to_num_replies
fo.num_topics = fo_num_topics
fo.num_posts = fo_num_posts
ThreadLocalORMSession.flush_all()
def import_news(project, pid, frs_mapping, sf_project_shortname, nbhd):
from forgeblog import model as BM
posts = os.listdir(os.path.join(options.output_dir, pid, 'news'))
if len(posts):
news_app = project.app_instance('news')
if not news_app:
news_app = project.install_app('blog', 'news', mount_label='News')
h.set_context(project.shortname, 'news', neighborhood=nbhd)
# make all the blog posts
for post in posts:
if '.json' == post[-5:]:
post_data = loadjson(pid, 'news', post)
create_date = datetime.strptime(post_data.createdOn, '%Y-%m-%d %H:%M:%S')
p = BM.BlogPost.query.get(title=post_data.title,
timestamp=create_date,
app_config_id=news_app.config._id)
if not p:
p = BM.BlogPost(title=post_data.title,
timestamp=create_date,
app_config_id=news_app.config._id)
p.text = convert_post_content(frs_mapping, sf_project_shortname, post_data.body, nbhd)
p.mod_date = create_date
p.state = 'published'
if not p.slug:
p.make_slug()
if not p.history().first():
p.commit()
ThreadLocalORMSession.flush_all()
M.Thread.new(discussion_id=p.app_config.discussion_id,
ref_id=p.index_id(),
subject='%s discussion' % p.title)
user = get_user(post_data.createdByUsername)
p.history().first().author=dict(
id=user._id,
username=user.username,
display_name=user.get_pref('display_name'))
ThreadLocalORMSession.flush_all()
def check_unsupported_tools(project):
docs = make_client(options.api_url, 'DocumentApp')
doc_count = 0
for doc in docs.service.getDocumentFolderList(s, project.id, recursive=True).dataRows:
if doc.title == 'Root Folder':
continue
doc_count += 1
if doc_count:
log.warn('Migrating documents is not supported, but found %s docs' % doc_count)
scm = make_client(options.api_url, 'ScmApp')
for repo in scm.service.getRepositoryList(s, project.id).dataRows:
log.warn('Migrating SCM repos is not supported, but found %s' % repo.repositoryPath)
tasks = make_client(options.api_url, 'TaskApp')
task_count = len(tasks.service.getTaskList(s, project.id, filters=None).dataRows)
if task_count:
log.warn('Migrating tasks is not supported, but found %s tasks' % task_count)
tracker = make_client(options.api_url, 'TrackerApp')
tracker_count = len(tracker.service.getArtifactList(s, project.id, filters=None).dataRows)
if tracker_count:
log.warn('Migrating trackers is not supported, but found %s tracker artifacts' % task_count)
def load(project_id, *paths):
in_file = os.path.join(options.output_dir, project_id, *paths)
with open(in_file) as input:
content = input.read()
return unicode(content, 'utf-8')
def loadjson(*args):
# Object for attribute access
return json.loads(load(*args), object_hook=Object)
def save(content, project, *paths):
out_file = os.path.join(options.output_dir, project.id, *paths)
if not os.path.exists(os.path.dirname(out_file)):
os.makedirs(os.path.dirname(out_file))
with open(out_file, 'w') as out:
out.write(content.encode('utf-8'))
def download_file(tool, url_path, *filepaths):
if tool == 'wiki':
action = 'viewAttachment'
elif tool == 'frs':
action = 'downloadFile'
else:
raise ValueError('tool %s not supported' % tool)
action_url = options.attachment_url % (tool, action)
out_file = os.path.join(options.output_dir, *filepaths)
if not os.path.exists(os.path.dirname(out_file)):
os.makedirs(os.path.dirname(out_file))
if '://' in url_path:
url = url_path
else:
hostname = urlparse(options.api_url).hostname
scheme = urlparse(options.api_url).scheme
url = scheme + '://' + hostname + action_url + urllib.quote(url_path)
log.debug('fetching %s' % url)
resp = loggedInOpener.open(url)
# if not logged in and this is private, you will get an html response instead of the file
# log in to make sure the file should really be html
if resp.headers.type == 'text/html':
# log in and save the file
resp = loggedInOpener.open(scheme + '://' + hostname + "/sf/sfmain/do/login", urllib.urlencode({
'username': options.username,
'password': options.password,
'returnToUrl': url,
'sfsubmit': 'submit'
}))
with open(out_file, 'w') as out:
out.write(resp.fp.read())
return out_file
bracket_macro = re.compile(r'\[(.*?)\]')
h1 = re.compile(r'^!!!', re.MULTILINE)
h2 = re.compile(r'^!!', re.MULTILINE)
h3 = re.compile(r'^!', re.MULTILINE)
re_stats = re.compile(r'#+ .* [Ss]tatistics\n+(.*\[sf:.*?Statistics\].*)+')
def wiki2markdown(markup):
'''
Partial implementation of http://help.collab.net/index.jsp?topic=/teamforge520/reference/wiki-wikisyntax.html
TODO: __ for bold
TODO: quote filenames with spaces, e.g. [[img src="foo bar.jpg"]]
'''
def bracket_handler(matchobj):
snippet = matchobj.group(1)
ext = snippet.rsplit('.')[-1].lower()
# TODO: support [foo|bar.jpg]
if snippet.startswith('sf:'):
# can't handle these macros
return matchobj.group(0)
elif ext in ('jpg', 'gif', 'png'):
filename = snippet.split('/')[-1]
return '[[img src=%s]]' % filename
elif '|' in snippet:
text, link = snippet.split('|', 1)
return '[%s](%s)' % (text, link)
else:
# regular link
return '<%s>' % snippet
markup = bracket_macro.sub(bracket_handler, markup or '')
markup = h1.sub('#', markup)
markup = h2.sub('##', markup)
markup = h3.sub('###', markup)
markup = re_stats.sub('', markup)
return markup
re_rel = re.compile(r'\b(rel\d+)\b')
def convert_post_content(frs_mapping, sf_project_shortname, text, nbhd):
def rel_handler(matchobj):
relno = matchobj.group(1)
path = frs_mapping.get(relno)
if path:
return '<a href="/projects/%s.%s/files/%s">%s</a>' % (
sf_project_shortname, nbhd.url_prefix.strip('/'), path, path)
else:
return relno
text = re_rel.sub(rel_handler, text or '')
return text
def find_image_references(markup):
'yields filenames'
for matchobj in bracket_macro.finditer(markup):
snippet = matchobj.group(1)
ext = snippet.rsplit('.')[-1].lower()
if ext in ('jpg', 'gif', 'png'):
yield snippet
def get_news(project):
'''
Extracts news posts
'''
app = make_client(options.api_url, 'NewsApp')
# find the forums
posts = app.service.getNewsPostList(s, project.id)
for post in posts.dataRows:
save(json.dumps(dict(post), default=str), project, 'news', post.id+'.json')
save_user(post.createdByUsername)
def get_discussion(project):
'''
Extracts discussion forums and posts
'''
app = make_client(options.api_url, 'DiscussionApp')
# find the forums
forums = app.service.getForumList(s, project.id)
for forum in forums.dataRows:
forumname = forum.path.split('.')[-1]
log.info('Retrieving data for forum: %s' % forumname)
save(json.dumps(dict(forum), default=str), project, 'forum', forumname+'.json')
# topic in this forum
topics = app.service.getTopicList(s, forum.id)
for topic in topics.dataRows:
save(json.dumps(dict(topic), default=str), project, 'forum', forumname, topic.id+'.json')
# posts in this topic
posts = app.service.getPostList(s, topic.id)
for post in posts.dataRows:
save(json.dumps(dict(post), default=str), project, 'forum', forumname, topic.id, post.id+'.json')
save_user(post.createdByUserName)
def get_homepage_wiki(project):
'''
Extracts home page and wiki pages
'''
wiki = make_client(options.api_url, 'WikiApp')
pages = {}
wiki_pages = wiki.service.getWikiPageList(s, project.id)
for wiki_page in wiki_pages.dataRows:
wiki_page = wiki.service.getWikiPageData(s, wiki_page.id)
pagename = wiki_page.path.split('/')[-1]
save(json.dumps(dict(wiki_page), default=str), project, 'wiki', pagename+'.json')
if not wiki_page.wikiText:
log.debug('skip blank wiki page %s' % wiki_page.path)
continue
pages[pagename] = wiki_page.wikiText
# PageApp does not provide a useful way to determine the Project Home special wiki page
# so use some heuristics
homepage = None
if '$ProjectHome' in pages and options.default_wiki_text not in pages['$ProjectHome']:
homepage = pages.pop('$ProjectHome')
elif 'HomePage' in pages and options.default_wiki_text not in pages['HomePage']:
homepage = pages.pop('HomePage')
elif '$ProjectHome' in pages:
homepage = pages.pop('$ProjectHome')
elif 'HomePage' in pages:
homepage = pages.pop('HomePage')
else:
log.warn('did not find homepage')
if homepage:
save(homepage, project, 'wiki', 'homepage_text.markdown')
for img_ref in find_image_references(homepage):
filename = img_ref.split('/')[-1]
if '://' in img_ref:
img_url = img_ref
else:
img_url = project.path + '/wiki/' + img_ref
download_file('wiki', img_url, project.id, 'wiki', 'homepage', filename)
for path, text in pages.iteritems():
if options.default_wiki_text in text:
log.debug('skipping default wiki page %s' % path)
else:
save(text, project, 'wiki', path+'.markdown')
for img_ref in find_image_references(text):
filename = img_ref.split('/')[-1]
if '://' in img_ref:
img_url = img_ref
else:
img_url = project.path + '/wiki/' + img_ref
download_file('wiki', img_url, project.id, 'wiki', path, filename)
def _dir_sql(created_on, project, dir_name, rel_path):
assert options.neighborhood_shortname
if not rel_path:
parent_directory = "'1'"
else:
parent_directory = "(SELECT pfs_path FROM pfs_path WHERE path_name = '%s/')" % rel_path
sql = """
UPDATE pfs
SET file_crtime = '%s'
WHERE source_pk = (SELECT project.project FROM project WHERE project.project_name = '%s.%s')
AND source_table = 'project'
AND pfs_type = 'd'
AND pfs_name = '%s'
AND parent_directory = %s;
""" % (created_on, convert_project_shortname(project.path), options.neighborhood_shortname, dir_name, parent_directory)
return sql
def get_files(project):
frs = make_client(options.api_url, 'FrsApp')
valid_pfs_filename = re.compile(r'(?![. ])[-_ +.,=#~@!()\[\]a-zA-Z0-9]+(?<! )$')
pfs_output_dir = os.path.join(os.path.abspath(options.output_dir), 'PFS', convert_project_shortname(project.path))
sql_updates = ''
def handle_path(obj, prev_path):
path_component = obj.title.strip().replace('/', ' ').replace('&','').replace(':','')
path = os.path.join(prev_path, path_component)
if not valid_pfs_filename.match(path_component):
log.error('Invalid filename: "%s"' % path)
save(json.dumps(dict(obj), default=str),
project, 'frs', path+'.json')
return path
frs_mapping = {}
for pkg in frs.service.getPackageList(s, project.id).dataRows:
pkg_path = handle_path(pkg, '')
pkg_details = frs.service.getPackageData(s, pkg.id) # download count
save(json.dumps(dict(pkg_details), default=str),
project, 'frs', pkg_path+'_details.json')
for rel in frs.service.getReleaseList(s, pkg.id).dataRows:
rel_path = handle_path(rel, pkg_path)
frs_mapping[rel['id']] = rel_path
rel_details = frs.service.getReleaseData(s, rel.id) # download count
save(json.dumps(dict(rel_details), default=str),
project, 'frs', rel_path+'_details.json')
for file in frs.service.getFrsFileList(s, rel.id).dataRows:
details = frs.service.getFrsFileData(s, file.id)
file_path = handle_path(file, rel_path)
save(json.dumps(dict(file,
lastModifiedBy=details.lastModifiedBy,
lastModifiedDate=details.lastModifiedDate,
),
default=str),
project,
'frs',
file_path+'.json'
)
if not options.skip_frs_download:
download_file('frs', rel.path + '/' + file.id, pfs_output_dir, file_path)
mtime = int(mktime(details.lastModifiedDate.timetuple()))
os.utime(os.path.join(pfs_output_dir, file_path), (mtime, mtime))
# releases
created_on = int(mktime(rel.createdOn.timetuple()))
mtime = int(mktime(rel.lastModifiedOn.timetuple()))
if os.path.exists(os.path.join(pfs_output_dir, rel_path)):
os.utime(os.path.join(pfs_output_dir, rel_path), (mtime, mtime))
sql_updates += _dir_sql(created_on, project, rel.title.strip(), pkg_path)
# packages
created_on = int(mktime(pkg.createdOn.timetuple()))
mtime = int(mktime(pkg.lastModifiedOn.timetuple()))
if os.path.exists(os.path.join(pfs_output_dir, pkg_path)):
os.utime(os.path.join(pfs_output_dir, pkg_path), (mtime, mtime))
sql_updates += _dir_sql(created_on, project, pkg.title.strip(), '')
# save pfs update sql for this project
with open(os.path.join(options.output_dir, 'pfs_updates.sql'), 'a') as out:
out.write('/* %s */' % project.id)
out.write(sql_updates)
save(json.dumps(frs_mapping), project, 'frs_mapping.json')
def get_parser(defaults):
optparser = OptionParser(
usage=('%prog [--options] [projID projID projID]\n'
'If no project ids are given, all projects will be migrated'))
optparser.set_defaults(**defaults)
# Command-line-only options
optparser.add_option(
'--extract-only', action='store_true', dest='extract',
help='Store data from the TeamForge API on the local filesystem; not load into Allura')
optparser.add_option(
'--load-only', action='store_true', dest='load',
help='Load into Allura previously-extracted data')
optparser.add_option(
'--config-file', dest='config_file',
help='Load options from config file')
# Command-line options with defaults in config file
optparser.add_option(
'--api-url', dest='api_url', help='e.g. https://hostname/ce-soap50/services/')
optparser.add_option(
'--attachment-url', dest='attachment_url')
optparser.add_option(
'--default-wiki-text', dest='default_wiki_text',
help='used in determining if a wiki page text is default or changed')
optparser.add_option(
'-u', '--username', dest='username')
optparser.add_option(
'-p', '--password', dest='password')
optparser.add_option(
'-o', '--output-dir', dest='output_dir')
optparser.add_option(
'--list-project-ids', action='store_true', dest='list_project_ids')
optparser.add_option(
'-n', '--neighborhood', dest='neighborhood',
help='Neighborhood full name, to load in to')
optparser.add_option(
'--n-shortname', dest='neighborhood_shortname',
help='Neighborhood shortname, for PFS extract SQL')
optparser.add_option(
'--skip-thread-import-id-when-reloading', action='store_true',
dest='skip_thread_import_id_when_reloading'
)
optparser.add_option(
'--skip-frs-download', action='store_true', dest='skip_frs_download')
optparser.add_option(
'--skip-wiki', action='store_true', dest='skip_wiki')
optparser.add_option(
'--skip-unsupported-check', action='store_true', dest='skip_unsupported_check')
return optparser
re_username = re.compile(r"^[a-z\-0-9]+$")
def make_valid_sf_username(orig_username):
sf_username = orig_username.replace('_','-').lower()
# FIXME username translation is hardcoded here:
sf_username = dict(
rlevy = 'ramilevy',
mkeisler = 'mkeisler',
bthale = 'bthale',
mmuller = 'mattjustmull',
MalcolmDwyer = 'slagheap',
tjyang = 'tjyang',
manaic = 'maniac76',
srinid = 'cnudav',
es = 'est016',
david_peyer = 'david-mmi',
okruse = 'ottokruse',
jvp = 'jvpmoto',
dmorelli = 'dmorelli',
).get(sf_username, sf_username + '-mmi')
if not re_username.match(sf_username):
adjusted_username = ''.join(
ch for ch in sf_username[:-4]
if ch.isalnum() or ch == '-') + '-mmi'
log.error('invalid sf_username characters: %s Changing it to %s',
sf_username, adjusted_username)
sf_username = adjusted_username
if len(sf_username) > 15:
adjusted_username = sf_username[0:15-4] + '-mmi'
log.error('invalid sf_username length: %s Changing it to %s',
sf_username, adjusted_username)
sf_username = adjusted_username
return sf_username
if __name__ == '__main__':
logging.basicConfig(level=logging.WARN)
log.setLevel(logging.DEBUG)
main()
def test_make_valid_sf_username():
tests = {
# basic
'foo':'foo-mmi',
# lookup
'rlevy':'ramilevy',
# too long
'u012345678901234567890': 'u0123456789-mmi',
'foo^213': 'foo213-mmi'
}
for k,v in tests.iteritems():
assert make_valid_sf_username(k) == v
def test_convert_post_content():
nbhd = Object()
nbhd.url_prefix = '/motorola/'
text = '''rel100? or ?rel101 or rel102 or rel103a or rel104'''
mapping = dict(
rel100='rel/100/',
rel101='rel/101/',
rel102='rel/102/',
rel103='rel/103/',
rel104='rel/104/')
converted = convert_post_content(mapping, 'foo', text, nbhd)
assert 'href="/projects/foo.motorola/files/rel/100' in converted, converted
assert 'href="/projects/foo.motorola/files/rel/101' in converted, converted
assert 'href="/projects/foo.motorola/files/rel/102' in converted, converted
assert 'href="/projects/foo.motorola/files/rel/103' not in converted, converted
assert 'href="/projects/foo.motorola/files/rel/104' in converted, converted
def test_convert_markup():
markup = '''
!this is the first headline
Please note that this project is for distributing, discussing, and supporting the open source software we release.
[http://www.google.com]
[SourceForge |http://www.sf.net]
[$ProjectHome/myimage.jpg]
[$ProjectHome/anotherimage.jpg]
!!! Project Statistics
|[sf:frsStatistics]|[sf:artifactStatistics]|
'''
new_markup = wiki2markdown(markup)
assert '\n[[img src=myimage.jpg]]\n[[img src=anotherimage.jpg]]\n' in new_markup
assert '\n###this is the first' in new_markup
assert '<http://www.google.com>' in new_markup
assert '[SourceForge ](http://www.sf.net)' in new_markup
assert '\n# Project Statistics' not in new_markup
assert '[sf:frsStatistics]' not in new_markup