api.py 31 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917
  1. #!/usr/bin/env python
  2. # Licensed to Cloudera, Inc. under one
  3. # or more contributor license agreements. See the NOTICE file
  4. # distributed with this work for additional information
  5. # regarding copyright ownership. Cloudera, Inc. licenses this file
  6. # to you under the Apache License, Version 2.0 (the
  7. # "License"); you may not use this file except in compliance
  8. # with the License. You may obtain a copy of the License at
  9. #
  10. # http://www.apache.org/licenses/LICENSE-2.0
  11. #
  12. # Unless required by applicable law or agreed to in writing, software
  13. # distributed under the License is distributed on an "AS IS" BASIS,
  14. # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  15. # See the License for the specific language governing permissions and
  16. # limitations under the License.
  17. from future import standard_library
  18. standard_library.install_aliases()
  19. import json
  20. import logging
  21. import sqlparse
  22. import sys
  23. from django.urls import reverse
  24. from django.db.models import Q
  25. from django.utils.translation import ugettext as _
  26. from django.views.decorators.http import require_GET, require_POST
  27. import opentracing.tracer
  28. from azure.abfs.__init__ import abfspath
  29. from desktop.conf import TASK_SERVER
  30. from desktop.lib.i18n import smart_str
  31. from desktop.lib.django_util import JsonResponse
  32. from desktop.models import Document2, Document, __paginate
  33. from indexer.file_format import HiveFormat
  34. from indexer.fields import Field
  35. from notebook.connectors.base import Notebook, QueryExpired, SessionExpired, QueryError, _get_snippet_name
  36. from notebook.connectors.hiveserver2 import HS2Api
  37. from notebook.decorators import api_error_handler, check_document_access_permission, check_document_modify_permission
  38. from notebook.models import escape_rows, make_notebook, upgrade_session_properties, get_api
  39. if sys.version_info[0] > 2:
  40. from urllib.parse import unquote as urllib_unquote
  41. else:
  42. from urllib import unquote as urllib_unquote
  43. LOG = logging.getLogger(__name__)
  44. DEFAULT_HISTORY_NAME = ''
  45. @require_POST
  46. @api_error_handler
  47. def create_notebook(request):
  48. response = {'status': -1}
  49. editor_type = request.POST.get('type', 'notebook')
  50. gist_id = request.POST.get('gist')
  51. directory_uuid = request.POST.get('directory_uuid')
  52. if gist_id:
  53. editor_type = 'impala'
  54. editor = make_notebook(
  55. name='name',
  56. description='desc',
  57. editor_type=editor_type,
  58. statement='SELECT ...',
  59. )
  60. else:
  61. editor = Notebook()
  62. data = editor.get_data()
  63. if editor_type != 'notebook':
  64. data['name'] = ''
  65. data['type'] = 'query-%s' % editor_type # TODO: Add handling for non-SQL types
  66. data['directoryUuid'] = directory_uuid
  67. editor.data = json.dumps(data)
  68. response['notebook'] = editor.get_data()
  69. response['status'] = 0
  70. return JsonResponse(response)
  71. @require_POST
  72. @check_document_access_permission
  73. @api_error_handler
  74. def create_session(request):
  75. response = {'status': -1}
  76. session = json.loads(request.POST.get('session', '{}'))
  77. properties = session.get('properties', [])
  78. response['session'] = get_api(request, session).create_session(lang=session['type'], properties=properties)
  79. response['status'] = 0
  80. return JsonResponse(response)
  81. @require_POST
  82. @check_document_access_permission
  83. @api_error_handler
  84. def close_session(request):
  85. response = {'status': -1}
  86. session = json.loads(request.POST.get('session', '{}'))
  87. response['session'] = get_api(request, {'type': session['type']}).close_session(session=session)
  88. response['status'] = 0
  89. return JsonResponse(response)
  90. def _execute_notebook(request, notebook, snippet):
  91. response = {'status': -1}
  92. result = None
  93. history = None
  94. historify = (notebook['type'] != 'notebook' or snippet.get('wasBatchExecuted')) and not notebook.get('skipHistorify')
  95. try:
  96. try:
  97. session = notebook.get('sessions') and notebook['sessions'][0] # Session reference for snippet execution without persisting it
  98. if historify:
  99. history = _historify(notebook, request.user)
  100. notebook = Notebook(document=history).get_data()
  101. interpreter = get_api(request, snippet)
  102. if snippet.get('interface') == 'sqlalchemy':
  103. interpreter.options['session'] = session
  104. with opentracing.tracer.start_span('interpreter') as span:
  105. response['handle'] = interpreter.execute(notebook, snippet)
  106. # Retrieve and remove the result from the handle
  107. if response['handle'].get('sync'):
  108. result = response['handle'].pop('result')
  109. finally:
  110. if historify:
  111. _snippet = [s for s in notebook['snippets'] if s['id'] == snippet['id']][0]
  112. if 'handle' in response: # No failure
  113. _snippet['result']['handle'] = response['handle']
  114. _snippet['result']['statements_count'] = response['handle'].get('statements_count', 1)
  115. _snippet['result']['statement_id'] = response['handle'].get('statement_id', 0)
  116. _snippet['result']['handle']['statement'] = response['handle'].get('statement', snippet['statement']).strip() # For non HS2, as non multi query yet
  117. else:
  118. _snippet['status'] = 'failed'
  119. if history: # If _historify failed, history will be None. If we get Atomic block exception, something underneath interpreter.execute() crashed and is not handled.
  120. history.update_data(notebook)
  121. history.save()
  122. response['history_id'] = history.id
  123. response['history_uuid'] = history.uuid
  124. if notebook['isSaved']: # Keep track of history of saved queries
  125. response['history_parent_uuid'] = history.dependencies.filter(type__startswith='query-').latest('last_modified').uuid
  126. except QueryError as ex: # We inject the history information from _historify() to the failed queries
  127. if response.get('history_id'):
  128. ex.extra['history_id'] = response['history_id']
  129. if response.get('history_uuid'):
  130. ex.extra['history_uuid'] = response['history_uuid']
  131. if response.get('history_parent_uuid'):
  132. ex.extra['history_parent_uuid'] = response['history_parent_uuid']
  133. raise ex
  134. # Inject and HTML escape results
  135. if result is not None:
  136. response['result'] = result
  137. response['result']['data'] = escape_rows(result['data'])
  138. response['status'] = 0
  139. return response
  140. @require_POST
  141. @check_document_access_permission
  142. @api_error_handler
  143. def execute(request, engine=None):
  144. notebook = json.loads(request.POST.get('notebook', '{}'))
  145. snippet = json.loads(request.POST.get('snippet', '{}'))
  146. with opentracing.tracer.start_span('notebook-execute') as span:
  147. span.set_tag('user-id', request.user.username)
  148. response = _execute_notebook(request, notebook, snippet)
  149. span.set_tag(
  150. 'query-id',
  151. response['handle']['guid'] if response.get('handle') and response['handle'].get('guid') else None
  152. )
  153. return JsonResponse(response)
  154. @require_POST
  155. @check_document_access_permission
  156. @api_error_handler
  157. def check_status(request):
  158. response = {'status': -1}
  159. notebook = json.loads(request.POST.get('notebook', '{}'))
  160. snippet = json.loads(request.POST.get('snippet', '{}'))
  161. if not snippet:
  162. nb_doc = Document2.objects.get_by_uuid(user=request.user, uuid=notebook['id'])
  163. notebook = Notebook(document=nb_doc).get_data()
  164. snippet = notebook['snippets'][0]
  165. try:
  166. with opentracing.tracer.start_span('notebook-check_status') as span:
  167. span.set_tag('user-id', request.user.username)
  168. span.set_tag(
  169. 'query-id',
  170. snippet['result']['handle']['guid'] if snippet['result'].get('handle') and snippet['result']['handle'].get('guid') else None
  171. )
  172. response['query_status'] = get_api(request, snippet).check_status(notebook, snippet)
  173. response['status'] = 0
  174. except SessionExpired:
  175. response['status'] = 'expired'
  176. raise
  177. except QueryExpired:
  178. response['status'] = 'expired'
  179. raise
  180. finally:
  181. if response['status'] == 0 and snippet['status'] != response['query_status']:
  182. status = response['query_status']['status']
  183. elif response['status'] == 'expired':
  184. status = 'expired'
  185. else:
  186. status = 'failed'
  187. if notebook['type'].startswith('query') or notebook.get('isManaged'):
  188. nb_doc = Document2.objects.get(id=notebook['id'])
  189. if nb_doc.can_write(request.user):
  190. nb = Notebook(document=nb_doc).get_data()
  191. if status != nb['snippets'][0]['status']:
  192. nb['snippets'][0]['status'] = status
  193. nb_doc.update_data(nb)
  194. nb_doc.save()
  195. return JsonResponse(response)
  196. @require_POST
  197. @check_document_access_permission
  198. @api_error_handler
  199. def fetch_result_data(request):
  200. response = {'status': -1}
  201. notebook = json.loads(request.POST.get('notebook', '{}'))
  202. snippet = json.loads(request.POST.get('snippet', '{}'))
  203. rows = json.loads(request.POST.get('rows', '100'))
  204. start_over = json.loads(request.POST.get('startOver', 'false'))
  205. with opentracing.tracer.start_span('notebook-fetch_result_data') as span:
  206. response['result'] = get_api(request, snippet).fetch_result(notebook, snippet, rows, start_over)
  207. span.set_tag('user-id', request.user.username)
  208. span.set_tag(
  209. 'query-id',
  210. snippet['result']['handle']['guid'] if snippet['result'].get('handle') and snippet['result']['handle'].get('guid') else None
  211. )
  212. # Materialize and HTML escape results
  213. if response['result'].get('data') and response['result'].get('type') == 'table' and not response['result'].get('isEscaped'):
  214. response['result']['data'] = escape_rows(response['result']['data'])
  215. response['result']['isEscaped'] = True
  216. response['status'] = 0
  217. return JsonResponse(response)
  218. @require_POST
  219. @check_document_access_permission
  220. @api_error_handler
  221. def fetch_result_metadata(request):
  222. response = {'status': -1}
  223. notebook = json.loads(request.POST.get('notebook', '{}'))
  224. snippet = json.loads(request.POST.get('snippet', '{}'))
  225. with opentracing.tracer.start_span('notebook-fetch_result_metadata') as span:
  226. response['result'] = get_api(request, snippet).fetch_result_metadata(notebook, snippet)
  227. span.set_tag('user-id', request.user.username)
  228. span.set_tag(
  229. 'query-id',
  230. snippet['result']['handle']['guid'] if snippet['result'].get('handle') and snippet['result']['handle'].get('guid') else None
  231. )
  232. response['status'] = 0
  233. return JsonResponse(response)
  234. @require_POST
  235. @check_document_access_permission
  236. @api_error_handler
  237. def fetch_result_size(request):
  238. response = {'status': -1}
  239. notebook = json.loads(request.POST.get('notebook', '{}'))
  240. snippet = json.loads(request.POST.get('snippet', '{}'))
  241. with opentracing.tracer.start_span('notebook-fetch_result_size') as span:
  242. response['result'] = get_api(request, snippet).fetch_result_size(notebook, snippet)
  243. span.set_tag('user-id', request.user.username)
  244. span.set_tag(
  245. 'query-id',
  246. snippet['result']['handle']['guid'] if snippet['result'].get('handle') and snippet['result']['handle'].get('guid') else None
  247. )
  248. response['status'] = 0
  249. return JsonResponse(response)
  250. @require_POST
  251. @check_document_access_permission
  252. @api_error_handler
  253. def cancel_statement(request):
  254. response = {'status': -1}
  255. notebook = json.loads(request.POST.get('notebook', '{}'))
  256. nb_doc = Document2.objects.get_by_uuid(user=request.user, uuid=notebook['uuid'])
  257. notebook = Notebook(document=nb_doc).get_data()
  258. snippet = notebook['snippets'][0]
  259. with opentracing.tracer.start_span('notebook-cancel_statement') as span:
  260. response['result'] = get_api(request, snippet).cancel(notebook, snippet)
  261. span.set_tag('user-id', request.user.username)
  262. span.set_tag(
  263. 'query-id',
  264. snippet['result']['handle']['guid'] if snippet['result'].get('handle') and snippet['result']['handle'].get('guid') else None
  265. )
  266. response['status'] = 0
  267. return JsonResponse(response)
  268. @require_POST
  269. @check_document_access_permission
  270. @api_error_handler
  271. def get_logs(request):
  272. response = {'status': -1}
  273. notebook = json.loads(request.POST.get('notebook', '{}'))
  274. snippet = json.loads(request.POST.get('snippet', '{}'))
  275. startFrom = request.POST.get('from')
  276. startFrom = int(startFrom) if startFrom else None
  277. size = request.POST.get('size')
  278. size = int(size) if size else None
  279. full_log = smart_str(request.POST.get('full_log', ''))
  280. db = get_api(request, snippet)
  281. with opentracing.tracer.start_span('notebook-get_logs') as span:
  282. logs = smart_str(db.get_log(notebook, snippet, startFrom=startFrom, size=size))
  283. span.set_tag('user-id', request.user.username)
  284. span.set_tag(
  285. 'query-id',
  286. snippet['result']['handle']['guid'] if snippet['result'].get('handle') and snippet['result']['handle'].get('guid') else None
  287. )
  288. full_log += logs
  289. jobs = db.get_jobs(notebook, snippet, full_log)
  290. response['logs'] = logs.strip()
  291. response['progress'] = min(db.progress(notebook, snippet, logs=full_log), 99) if snippet['status'] != 'available' and snippet['status'] != 'success' else 100
  292. response['jobs'] = jobs
  293. response['isFullLogs'] = db.get_log_is_full_log(notebook, snippet)
  294. response['status'] = 0
  295. return JsonResponse(response)
  296. def _save_notebook(notebook, user):
  297. notebook_type = notebook.get('type', 'notebook')
  298. save_as = False
  299. if notebook.get('parentSavedQueryUuid'): # We save into the original saved query, not into the query history
  300. notebook_doc = Document2.objects.get_by_uuid(user=user, uuid=notebook['parentSavedQueryUuid'])
  301. elif notebook.get('id'):
  302. notebook_doc = Document2.objects.get(id=notebook['id'])
  303. else:
  304. notebook_doc = Document2.objects.create(name=notebook['name'], uuid=notebook['uuid'], type=notebook_type, owner=user)
  305. Document.objects.link(notebook_doc, owner=notebook_doc.owner, name=notebook_doc.name, description=notebook_doc.description, extra=notebook_type)
  306. save_as = True
  307. if notebook.get('directoryUuid'):
  308. notebook_doc.parent_directory = Document2.objects.get_by_uuid(user=user, uuid=notebook.get('directoryUuid'), perm_type='write')
  309. else:
  310. notebook_doc.parent_directory = Document2.objects.get_home_directory(user)
  311. notebook['isSaved'] = True
  312. notebook['isHistory'] = False
  313. notebook['id'] = notebook_doc.id
  314. _clear_sessions(notebook)
  315. notebook_doc1 = notebook_doc._get_doc1(doc2_type=notebook_type)
  316. notebook_doc.update_data(notebook)
  317. notebook_doc.search = _get_statement(notebook)
  318. notebook_doc.name = notebook_doc1.name = notebook['name']
  319. notebook_doc.description = notebook_doc1.description = notebook['description']
  320. notebook_doc.save()
  321. notebook_doc1.save()
  322. return notebook_doc, save_as
  323. @api_error_handler
  324. @require_POST
  325. @check_document_modify_permission()
  326. def save_notebook(request):
  327. response = {'status': -1}
  328. notebook = json.loads(request.POST.get('notebook', '{}'))
  329. notebook_doc, save_as = _save_notebook(notebook, request.user)
  330. response['status'] = 0
  331. response['save_as'] = save_as
  332. response.update(notebook_doc.to_dict())
  333. response['message'] = request.POST.get('editorMode') == 'true' and _('Query saved successfully') or _('Notebook saved successfully')
  334. return JsonResponse(response)
  335. def _clear_sessions(notebook):
  336. notebook['sessions'] = [_s for _s in notebook['sessions'] if _s['type'] in ('scala', 'spark', 'pyspark', 'sparkr', 'r')]
  337. def _historify(notebook, user):
  338. query_type = notebook['type']
  339. name = notebook['name'] if (notebook['name'] and notebook['name'].strip() != '') else DEFAULT_HISTORY_NAME
  340. is_managed = notebook.get('isManaged') == True # Prevents None
  341. if is_managed and Document2.objects.filter(uuid=notebook['uuid']).exists():
  342. history_doc = Document2.objects.get(uuid=notebook['uuid'])
  343. else:
  344. history_doc = Document2.objects.create(
  345. name=name,
  346. type=query_type,
  347. owner=user,
  348. is_history=True,
  349. is_managed=is_managed
  350. )
  351. # Link history of saved query
  352. if notebook['isSaved']:
  353. parent_doc = Document2.objects.get(uuid=notebook.get('parentSavedQueryUuid') or notebook['uuid']) # From previous history query or initial saved query
  354. notebook['parentSavedQueryUuid'] = parent_doc.uuid
  355. history_doc.dependencies.add(parent_doc)
  356. if not is_managed:
  357. Document.objects.link(
  358. history_doc,
  359. name=history_doc.name,
  360. owner=history_doc.owner,
  361. description=history_doc.description,
  362. extra=query_type
  363. )
  364. notebook['uuid'] = history_doc.uuid
  365. _clear_sessions(notebook)
  366. history_doc.update_data(notebook)
  367. history_doc.search = _get_statement(notebook)
  368. history_doc.save()
  369. return history_doc
  370. def _get_statement(notebook):
  371. if notebook['snippets'] and len(notebook['snippets']) > 0:
  372. return Notebook.statement_with_variables(notebook['snippets'][0])
  373. return ''
  374. @require_GET
  375. @api_error_handler
  376. @check_document_access_permission
  377. def get_history(request):
  378. response = {'status': -1}
  379. doc_type = request.GET.get('doc_type')
  380. doc_text = request.GET.get('doc_text')
  381. page = min(int(request.GET.get('page', 1)), 100)
  382. limit = min(int(request.GET.get('limit', 50)), 100)
  383. is_notification_manager = request.GET.get('is_notification_manager', 'false') == 'true'
  384. if is_notification_manager:
  385. docs = Document2.objects.get_tasks_history(user=request.user)
  386. else:
  387. docs = Document2.objects.get_history(doc_type='query-%s' % doc_type, user=request.user)
  388. if doc_text:
  389. docs = docs.filter(Q(name__icontains=doc_text) | Q(description__icontains=doc_text) | Q(search__icontains=doc_text))
  390. # Paginate
  391. docs = docs.order_by('-last_modified')
  392. response['count'] = docs.count()
  393. docs = __paginate(page, limit, queryset=docs)['documents']
  394. history = []
  395. for doc in docs:
  396. notebook = Notebook(document=doc).get_data()
  397. if 'snippets' in notebook:
  398. statement = notebook['description'] if is_notification_manager else _get_statement(notebook)
  399. history.append({
  400. 'name': doc.name,
  401. 'id': doc.id,
  402. 'uuid': doc.uuid,
  403. 'type': doc.type,
  404. 'data': {
  405. 'statement': statement[:1001] if statement else '',
  406. 'lastExecuted': notebook['snippets'][0].get('lastExecuted', -1),
  407. 'status': notebook['snippets'][0]['status'],
  408. 'parentSavedQueryUuid': notebook.get('parentSavedQueryUuid', '')
  409. } if notebook['snippets'] else {},
  410. 'absoluteUrl': doc.get_absolute_url(),
  411. })
  412. else:
  413. LOG.error('Incomplete History Notebook: %s' % notebook)
  414. response['history'] = sorted(history, key=lambda row: row['data']['lastExecuted'], reverse=True)
  415. response['message'] = _('History fetched')
  416. response['status'] = 0
  417. return JsonResponse(response)
  418. @require_POST
  419. @api_error_handler
  420. @check_document_modify_permission()
  421. def clear_history(request):
  422. response = {'status': -1}
  423. notebook = json.loads(request.POST.get('notebook'), '{}')
  424. doc_type = request.POST.get('doc_type')
  425. is_notification_manager = request.POST.get('is_notification_manager', 'false') == 'true'
  426. if is_notification_manager:
  427. history = Document2.objects.get_tasks_history(user=request.user)
  428. else:
  429. history = Document2.objects.get_history(doc_type='query-%s' % doc_type, user=request.user)
  430. response['updated'] = history.delete()
  431. response['message'] = _('History cleared !')
  432. response['status'] = 0
  433. return JsonResponse(response)
  434. @require_GET
  435. @check_document_access_permission
  436. def open_notebook(request):
  437. response = {'status': -1}
  438. notebook_id = request.GET.get('notebook')
  439. notebook = Notebook(document=Document2.objects.get(id=notebook_id))
  440. notebook = upgrade_session_properties(request, notebook)
  441. response['status'] = 0
  442. response['notebook'] = notebook.get_json()
  443. response['message'] = _('Notebook loaded successfully')
  444. @require_POST
  445. @check_document_access_permission
  446. def close_notebook(request):
  447. response = {'status': -1, 'result': []}
  448. notebook = json.loads(request.POST.get('notebook', '{}'))
  449. for session in [_s for _s in notebook['sessions'] if _s['type'] in ('scala', 'spark', 'pyspark', 'sparkr', 'r')]:
  450. try:
  451. response['result'].append(get_api(request, session).close_session(session))
  452. except QueryExpired:
  453. pass
  454. except Exception as e:
  455. LOG.exception('Error closing session %s' % str(e))
  456. for snippet in [_s for _s in notebook['snippets'] if _s['type'] in ('hive', 'impala')]:
  457. try:
  458. if snippet['status'] != 'running':
  459. response['result'].append(get_api(request, snippet).close_statement(notebook, snippet))
  460. else:
  461. LOG.info('Not closing SQL snippet as still running.')
  462. except QueryExpired:
  463. pass
  464. except Exception as e:
  465. LOG.exception('Error closing statement %s' % str(e))
  466. response['status'] = 0
  467. response['message'] = _('Notebook closed successfully')
  468. return JsonResponse(response)
  469. @require_POST
  470. @check_document_access_permission
  471. def close_statement(request):
  472. response = {'status': -1}
  473. # Passed by check_document_access_permission but unused by APIs
  474. notebook = json.loads(request.POST.get('notebook', '{}'))
  475. nb_doc = Document2.objects.get_by_uuid(user=request.user, uuid=notebook['uuid'])
  476. notebook = Notebook(document=nb_doc).get_data()
  477. snippet = notebook['snippets'][0]
  478. try:
  479. with opentracing.tracer.start_span('notebook-close_statement') as span:
  480. response['result'] = get_api(request, snippet).close_statement(notebook, snippet)
  481. span.set_tag('user-id', request.user.username)
  482. span.set_tag(
  483. 'query-id',
  484. snippet['result']['handle']['guid'] if snippet['result'].get('handle') and snippet['result']['handle'].get('guid') else None
  485. )
  486. except QueryExpired:
  487. pass
  488. response['status'] = 0
  489. response['message'] = _('Statement closed !')
  490. return JsonResponse(response)
  491. @require_POST
  492. @check_document_access_permission
  493. @api_error_handler
  494. def autocomplete(request, server=None, database=None, table=None, column=None, nested=None):
  495. response = {'status': -1}
  496. # Passed by check_document_access_permission but unused by APIs
  497. notebook = json.loads(request.POST.get('notebook', '{}'))
  498. snippet = json.loads(request.POST.get('snippet', '{}'))
  499. try:
  500. autocomplete_data = get_api(request, snippet).autocomplete(snippet, database, table, column, nested)
  501. response.update(autocomplete_data)
  502. except QueryExpired:
  503. pass
  504. response['status'] = 0
  505. return JsonResponse(response)
  506. @require_POST
  507. @check_document_access_permission
  508. @api_error_handler
  509. def get_sample_data(request, server=None, database=None, table=None, column=None):
  510. response = {'status': -1}
  511. # Passed by check_document_access_permission but unused by APIs
  512. notebook = json.loads(request.POST.get('notebook', '{}'))
  513. snippet = json.loads(request.POST.get('snippet', '{}'))
  514. async = json.loads(request.POST.get('async', 'false'))
  515. operation = json.loads(request.POST.get('operation', '"default"'))
  516. sample_data = get_api(request, snippet).get_sample_data(snippet, database, table, column, async=async, operation=operation)
  517. response.update(sample_data)
  518. response['status'] = 0
  519. return JsonResponse(response)
  520. @require_POST
  521. @check_document_access_permission
  522. @api_error_handler
  523. def explain(request):
  524. response = {'status': -1}
  525. notebook = json.loads(request.POST.get('notebook', '{}'))
  526. snippet = json.loads(request.POST.get('snippet', '{}'))
  527. response = get_api(request, snippet).explain(notebook, snippet)
  528. return JsonResponse(response)
  529. @require_POST
  530. @api_error_handler
  531. def format(request):
  532. response = {'status': 0}
  533. statements = request.POST.get('statements', '')
  534. response['formatted_statements'] = sqlparse.format(statements, reindent=True, keyword_case='upper') # SQL only currently
  535. return JsonResponse(response)
  536. @require_POST
  537. @check_document_access_permission
  538. @api_error_handler
  539. def export_result(request):
  540. response = {'status': -1, 'message': _('Success')}
  541. # Passed by check_document_access_permission but unused by APIs
  542. notebook = json.loads(request.POST.get('notebook', '{}'))
  543. snippet = json.loads(request.POST.get('snippet', '{}'))
  544. data_format = json.loads(request.POST.get('format', '"hdfs-file"'))
  545. destination = urllib_unquote(json.loads(request.POST.get('destination', '""')))
  546. overwrite = json.loads(request.POST.get('overwrite', 'false'))
  547. is_embedded = json.loads(request.POST.get('is_embedded', 'false'))
  548. start_time = json.loads(request.POST.get('start_time', '-1'))
  549. api = get_api(request, snippet)
  550. if data_format == 'hdfs-file': # Blocking operation, like downloading
  551. if request.fs.isdir(destination):
  552. if notebook.get('name'):
  553. destination += '/%(name)s.csv' % notebook
  554. else:
  555. destination += '/%(type)s-%(id)s.csv' % notebook
  556. if overwrite and request.fs.exists(destination):
  557. request.fs.do_as_user(request.user.username, request.fs.rmtree, destination)
  558. response['watch_url'] = api.export_data_as_hdfs_file(snippet, destination, overwrite)
  559. response['status'] = 0
  560. request.audit = {
  561. 'operation': 'EXPORT',
  562. 'operationText': 'User %s exported to HDFS destination: %s' % (request.user.username, destination),
  563. 'allowed': True
  564. }
  565. elif data_format == 'hive-table':
  566. if is_embedded:
  567. sql, success_url = api.export_data_as_table(notebook, snippet, destination)
  568. task = make_notebook(
  569. name=_('Export %s query to table %s') % (snippet['type'], destination),
  570. description=_('Query %s to %s') % (_get_snippet_name(notebook), success_url),
  571. editor_type=snippet['type'],
  572. statement=sql,
  573. status='ready',
  574. database=snippet['database'],
  575. on_success_url=success_url,
  576. last_executed=start_time,
  577. is_task=True
  578. )
  579. response = task.execute(request)
  580. else:
  581. notebook_id = notebook['id'] or request.GET.get('editor', request.GET.get('notebook'))
  582. response['watch_url'] = reverse('notebook:execute_and_watch') + '?action=save_as_table&notebook=' + str(notebook_id) + '&snippet=0&destination=' + destination
  583. response['status'] = 0
  584. request.audit = {
  585. 'operation': 'EXPORT',
  586. 'operationText': 'User %s exported to Hive table: %s' % (request.user.username, destination),
  587. 'allowed': True
  588. }
  589. elif data_format == 'hdfs-directory':
  590. if destination.lower().startswith("abfs"):
  591. destination = abfspath(destination)
  592. if is_embedded:
  593. sql, success_url = api.export_large_data_to_hdfs(notebook, snippet, destination)
  594. task = make_notebook(
  595. name=_('Export %s query to directory') % snippet['type'],
  596. description=_('Query %s to %s') % (_get_snippet_name(notebook), success_url),
  597. editor_type=snippet['type'],
  598. statement=sql,
  599. status='ready-execute',
  600. database=snippet['database'],
  601. on_success_url=success_url,
  602. last_executed=start_time,
  603. is_task=True
  604. )
  605. response = task.execute(request)
  606. else:
  607. notebook_id = notebook['id'] or request.GET.get('editor', request.GET.get('notebook'))
  608. response['watch_url'] = reverse('notebook:execute_and_watch') + '?action=insert_as_query&notebook=' + str(notebook_id) + '&snippet=0&destination=' + destination
  609. response['status'] = 0
  610. request.audit = {
  611. 'operation': 'EXPORT',
  612. 'operationText': 'User %s exported to HDFS directory: %s' % (request.user.username, destination),
  613. 'allowed': True
  614. }
  615. elif data_format in ('search-index', 'dashboard'):
  616. # Open the result in the Dashboard via a SQL sub-query or the Import wizard (quick vs scalable)
  617. if is_embedded:
  618. notebook_id = notebook['id'] or request.GET.get('editor', request.GET.get('notebook'))
  619. if data_format == 'dashboard':
  620. engine = notebook['type'].replace('query-', '')
  621. response['watch_url'] = reverse('dashboard:browse', kwargs={'name': notebook_id}) + '?source=query&engine=%(engine)s' % {'engine': engine}
  622. response['status'] = 0
  623. else:
  624. sample = get_api(request, snippet).fetch_result(notebook, snippet, rows=4, start_over=True)
  625. for col in sample['meta']:
  626. col['type'] = HiveFormat.FIELD_TYPE_TRANSLATE.get(col['type'], 'string')
  627. response['status'] = 0
  628. response['id'] = notebook_id
  629. response['name'] = _get_snippet_name(notebook)
  630. response['source_type'] = 'query'
  631. response['target_type'] = 'index'
  632. response['target_path'] = destination
  633. response['sample'] = list(sample['data'])
  634. response['columns'] = [
  635. Field(col['name'], col['type']).to_dict() for col in sample['meta']
  636. ]
  637. else:
  638. notebook_id = notebook['id'] or request.GET.get('editor', request.GET.get('notebook'))
  639. response['watch_url'] = reverse('notebook:execute_and_watch') + '?action=index_query&notebook=' + str(notebook_id) + '&snippet=0&destination=' + destination
  640. response['status'] = 0
  641. if response.get('status') != 0:
  642. response['message'] = _('Exporting result failed.')
  643. return JsonResponse(response)
  644. @require_POST
  645. @check_document_access_permission
  646. @api_error_handler
  647. def statement_risk(request):
  648. response = {'status': -1, 'message': ''}
  649. notebook = json.loads(request.POST.get('notebook', '{}'))
  650. snippet = json.loads(request.POST.get('snippet', '{}'))
  651. api = HS2Api(request.user, snippet)
  652. response['query_complexity'] = api.statement_risk(notebook, snippet)
  653. response['status'] = 0
  654. return JsonResponse(response)
  655. @require_POST
  656. @check_document_access_permission
  657. @api_error_handler
  658. def statement_compatibility(request):
  659. response = {'status': -1, 'message': ''}
  660. notebook = json.loads(request.POST.get('notebook', '{}'))
  661. snippet = json.loads(request.POST.get('snippet', '{}'))
  662. source_platform = request.POST.get('sourcePlatform')
  663. target_platform = request.POST.get('targetPlatform')
  664. api = get_api(request, snippet)
  665. response['query_compatibility'] = api.statement_compatibility(notebook, snippet, source_platform=source_platform, target_platform=target_platform)
  666. response['status'] = 0
  667. return JsonResponse(response)
  668. @require_POST
  669. @check_document_access_permission
  670. @api_error_handler
  671. def statement_similarity(request):
  672. response = {'status': -1, 'message': ''}
  673. notebook = json.loads(request.POST.get('notebook', '{}'))
  674. snippet = json.loads(request.POST.get('snippet', '{}'))
  675. source_platform = request.POST.get('sourcePlatform')
  676. api = get_api(request, snippet)
  677. response['statement_similarity'] = api.statement_similarity(notebook, snippet, source_platform=source_platform)
  678. response['status'] = 0
  679. return JsonResponse(response)
  680. @require_POST
  681. @check_document_access_permission
  682. @api_error_handler
  683. def get_external_statement(request):
  684. response = {'status': -1, 'message': ''}
  685. notebook = json.loads(request.POST.get('notebook', '{}'))
  686. snippet = json.loads(request.POST.get('snippet', '{}'))
  687. if snippet.get('statementType') == 'file':
  688. response['statement'] = _get_statement_from_file(request.user, request.fs, snippet)
  689. elif snippet.get('statementType') == 'document':
  690. notebook = Notebook(Document2.objects.get_by_uuid(user=request.user, uuid=snippet['associatedDocumentUuid'], perm_type='read'))
  691. response['statement'] = notebook.get_str()
  692. response['status'] = 0
  693. return JsonResponse(response)
  694. def _get_statement_from_file(user, fs, snippet):
  695. script_path = snippet['statementPath']
  696. if script_path:
  697. script_path = script_path.replace('hdfs://', '')
  698. if fs.do_as_user(user, fs.isfile, script_path):
  699. return fs.do_as_user(user, fs.read, script_path, 0, 16 * 1024 ** 2)
  700. @require_POST
  701. @api_error_handler
  702. def describe(request, database, table=None, column=None):
  703. response = {'status': -1, 'message': ''}
  704. notebook = json.loads(request.POST.get('notebook', '{}'))
  705. source_type = request.POST.get('source_type', '')
  706. snippet = {'type': source_type}
  707. describe = get_api(request, snippet).describe(notebook, snippet, database, table, column=column)
  708. response.update(describe)
  709. return JsonResponse(response)