def anonymize_table(table_id): if project.verbose: print('ANONYMIZE TABLE', project.task['bigquery']['to']['dataset'], table_id) schema = table_to_schema(project.task['auth'], project.task['bigquery']['from']['project'], project.task['bigquery']['from']['dataset'], table_id) rows = table_to_rows(project.task['auth'], project.task['bigquery']['from']['project'], project.task['bigquery']['from']['dataset'], table_id, as_object=True) rows = anonymize_rows(rows, project.task['bigquery']['to'].get('columns', [])) json_to_table(project.task['auth'], project.task['bigquery']['to']['project'], project.task['bigquery']['to']['dataset'], table_id, rows, schema, disposition='WRITE_TRUNCATE')
def anonymize_query(): if project.verbose: print('ANONYMIZE QUERY', project.task['bigquery']['from']['query']) schema = query_to_schema( project.task['auth'], project.task['bigquery']['from']['project'], project.task['bigquery']['from']['dataset'], project.task['bigquery']['from']['query'], legacy=project.task['bigquery']['from'].get('legacy', False), ) rows = query_to_rows(project.task['auth'], project.task['bigquery']['from']['project'], project.task['bigquery']['from']['dataset'], project.task['bigquery']['from']['query'], legacy=project.task['bigquery']['from'].get( 'legacy', False), as_object=True) rows = anonymize_rows(rows, project.task['bigquery']['to'].get('columns', [])) json_to_table(project.task['auth'], project.task['bigquery']['to']['project'], project.task['bigquery']['to']['dataset'], project.task['bigquery']['to']['table'], rows, schema, disposition='WRITE_TRUNCATE')
def bigquery_function(): """Generate custom tables or functions. See: scripts/bigquery_function.json """ if project.verbose: print('FUNCTION', project.task['function']) if project.task['function'] == 'Pearson Significance Test': run_query(project.task['auth'], project.id, pearson_significance_test(), False, project.task['to']['dataset']) elif project.task['function'] == 'US Geography': json_to_table(project.task['auth'], project.id, project.task['to']['dataset'], 'US_Geography', US_GEOGRAPHY_DATA, US_GEOGRAPHY_SCHEMA)
def anonymize_table(table_id): if project.verbose: print('ANONYMIZE TABLE', project.task['bigquery']['to']['dataset'], table_id) schema = API_BigQuery(project.task['auth']).tables().get( projectId=project.task['bigquery']['from']['project'], datasetId=project.task['bigquery']['from']['dataset'], tableId=table_id).execute()['schema']['fields'] is_object = any([s['type'] == 'RECORD' for s in schema]) rows = table_to_rows(project.task['auth'], project.task['bigquery']['from']['project'], project.task['bigquery']['from']['dataset'], table_id, as_object=is_object) rows = anonymize_rows(rows, schema, is_object) if is_object: json_to_table(project.task['auth'], project.task['bigquery']['to']['project'], project.task['bigquery']['to']['dataset'], table_id, rows, schema, disposition='WRITE_TRUNCATE') else: rows_to_table(project.task['auth'], project.task['bigquery']['to']['project'], project.task['bigquery']['to']['dataset'], table_id, rows, schema, skip_rows=0, disposition='WRITE_TRUNCATE')
def move_entity(project, path, table, schema, disposition): if 'prefix' in project.task: table = '%s_%s' % (project.task['prefix'], table) if 'out' in project.task: auth = project.task['out']['bigquery'].get('auth', project.task['auth']) dataset = project.task['out']['bigquery']['dataset'] # deprecated, need out to allow auth mix on in and out else: auth = project.task['auth'] dataset = project.task['dataset'] # read the entity file in parts records = get_entity(path) # write each part json_to_table( auth, project.id, dataset, table, records, schema=schema, disposition=disposition)
def put_rows(auth, destination, rows, variant=''): """Processes standard write JSON block for dynamic export of data. Allows us to quickly write the results of a script to a destination. For example write the results of a DCM report into BigQuery. - Will write to multiple destinations if specified. - Extensible, add a handler to define a new destination ( be kind update the documentation json ). Include the following JSON in a recipe, then in the run.py handler when encountering that block pass it to this function and use the returned results. from utils.data import put_rows var_json = { "out":{ "bigquery":{ "dataset": [ string ], "table": [ string ] "schema": [ json - standard bigquery schema json ], "skip_rows": [ integer - for removing header ] "disposition": [ string - same as BigQuery documentation ] }, "sheets":{ "sheet":[ string - full URL, suggest using share link ], "tab":[ string ], "range":[ string - A1:A notation ] "delete": [ boolean - if sheet range should be cleared before writing ] }, "storage":{ "bucket": [ string ], "path": [ string ] }, "file":[ string - full path to place to write file ] } } values = put_rows('user', var_json) Or you can use it directly with project singleton. from util.project import project from utils.data import put_rows @project.from_parameters def something(): values = get_rows(project.task['auth'], project.task['out']) if __name__ == "__main__": something() Args: auth: (string) The type of authentication to use, user or service. destination: (json) A json block resembling var_json described above. rows ( list ) The data being written as a list object. variant (string) Appended to destination to differentieate multiple objects Returns: If single_cell is False: Returns a list of row values [[v1], [v2], ... ] If single_cell is True: Returns a list of values [v1, v2, ...] """ if 'bigquery' in destination: if destination['bigquery'].get('format', 'CSV') == 'JSON': json_to_table( destination['bigquery'].get('auth', auth), destination['bigquery'].get('project_id', project.id), destination['bigquery']['dataset'], destination['bigquery']['table'] + variant, rows, destination['bigquery'].get('schema', []), destination['bigquery'].get('disposition', 'WRITE_TRUNCATE'), ) elif destination['bigquery'].get('is_incremental_load', False) == True: incremental_rows_to_table( destination['bigquery'].get('auth', auth), destination['bigquery'].get('project_id', project.id), destination['bigquery']['dataset'], destination['bigquery']['table'] + variant, rows, destination['bigquery'].get('schema', []), destination['bigquery'].get( 'skip_rows', 1), #0 if 'schema' in destination['bigquery'] else 1), destination['bigquery'].get('disposition', 'WRITE_APPEND'), billing_project_id=project.id) else: rows_to_table( destination['bigquery'].get('auth', auth), destination['bigquery'].get('project_id', project.id), destination['bigquery']['dataset'], destination['bigquery']['table'] + variant, rows, destination['bigquery'].get('schema', []), destination['bigquery'].get( 'skip_rows', 1), #0 if 'schema' in destination['bigquery'] else 1), destination['bigquery'].get('disposition', 'WRITE_TRUNCATE'), ) if 'sheets' in destination: if destination['sheets'].get('delete', False): sheets_clear( auth, destination['sheets']['sheet'], destination['sheets']['tab'] + variant, destination['sheets']['range'], ) sheets_write(auth, destination['sheets']['sheet'], destination['sheets']['tab'] + variant, destination['sheets']['range'], rows) if 'file' in destination: path_out, file_ext = destination['file'].rsplit('.', 1) file_out = path_out + variant + '.' + file_ext if project.verbose: print('SAVING', file_out) makedirs_safe(parse_path(file_out)) with open(file_out, 'w') as save_file: save_file.write(rows_to_csv(rows).read()) if 'storage' in destination and destination['storage'].get( 'bucket') and destination['storage'].get('path'): # create the bucket bucket_create(auth, project.id, destination['storage']['bucket']) # put the file file_out = destination['storage']['bucket'] + ':' + destination[ 'storage']['path'] + variant if project.verbose: print('SAVING', file_out) object_put(auth, file_out, rows_to_csv(rows)) if 'sftp' in destination: try: cnopts = pysftp.CnOpts() cnopts.hostkeys = None path_out, file_out = destination['sftp']['file'].rsplit('.', 1) file_out = path_out + variant + file_out sftp = pysftp.Connection(host=destination['sftp']['host'], username=destination['sftp']['username'], password=destination['sftp']['password'], port=destination['sftp']['port'], cnopts=cnopts) if '/' in file_out: dir_out, file_out = file_out.rsplit('/', 1) sftp.cwd(dir_out) sftp.putfo(rows_to_csv(rows), file_out) except e: print(str(e)) traceback.print_exc()
def put_rows(auth, destination, filename, rows, variant=''): """Processes standard write JSON block for dynamic export of data. Allows us to quickly write the results of a script to a destination. For example write the results of a DCM report into BigQuery. - Will write to multiple destinations if specified. - Extensible, add a handler to define a new destination ( be kind update the documentation json ). Include the following JSON in a recipe, then in the run.py handler when encountering that block pass it to this function and use the returned results. from utils.data import put_rows var_json = { "out":{ "bigquery":{ "dataset": [ string ], "table": [ string ] "schema": [ json - standard bigquery schema json ], "skip_rows": [ integer - for removing header ] "disposition": [ string - same as BigQuery documentation ] }, "sheets":{ "url":[ string - full URL, suggest using share link ], "tab":[ string ], "range":[ string - A1:A notation ] "delete": [ boolean - if sheet range should be cleared before writing ] }, "storage":{ "bucket": [ string ], "path": [ string ] }, "directory":[ string - full path to place to write file ] } } values = put_rows('user', var_json) Or you can use it directly with project singleton. from util.project import project from utils.data import put_rows @project.from_parameters def something(): values = get_rows(project.task['auth'], project.task['out']) if __name__ == "__main__": something() Args: auth: (string) The type of authentication to use, user or service. destination: (json) A json block resembling var_json described above. filename: (string) A unique filename if writing to medium requiring one, Usually gnerated by script. rows ( list ) The data being written as a list object. variant ( string ) Appends this to the destination name to create a variant ( for example when downloading multiple tabs in a sheet ). Returns: If single_cell is False: Returns a list of row values [[v1], [v2], ... ] If single_cell is True: Returns a list of values [v1, v2, ...] """ if 'bigquery' in destination: if destination['bigquery'].get('format', 'CSV') == 'JSON': json_to_table( destination['bigquery'].get('auth', auth), destination['bigquery'].get('project_id', project.id), destination['bigquery']['dataset'], destination['bigquery']['table'] + variant, rows, destination['bigquery'].get('schema', []), destination['bigquery'].get('disposition', 'WRITE_TRUNCATE'), ) elif destination['bigquery'].get('is_incremental_load', False) == True: incremental_rows_to_table( destination['bigquery'].get('auth', auth), destination['bigquery'].get('project_id', project.id), destination['bigquery']['dataset'], destination['bigquery']['table'] + variant, rows, destination['bigquery'].get('schema', []), destination['bigquery'].get( 'skip_rows', 1), #0 if 'schema' in destination['bigquery'] else 1), destination['bigquery'].get('disposition', 'WRITE_APPEND'), billing_project_id=project.id) else: rows_to_table( destination['bigquery'].get('auth', auth), destination['bigquery'].get('project_id', project.id), destination['bigquery']['dataset'], destination['bigquery']['table'] + variant, rows, destination['bigquery'].get('schema', []), destination['bigquery'].get( 'skip_rows', 1), #0 if 'schema' in destination['bigquery'] else 1), destination['bigquery'].get('disposition', 'WRITE_TRUNCATE'), ) if 'sheets' in destination: if destination['sheets'].get('delete', False): sheets_clear(auth, destination['sheets']['sheet'], destination['sheets']['tab'] + variant, destination['sheets']['range']) sheets_write(auth, destination['sheets']['sheet'], destination['sheets']['tab'] + variant, destination['sheets']['range'], rows) if 'directory' in destination: file_out = destination['directory'] + variant + filename if project.verbose: print 'SAVING', file_out makedirs_safe(parse_path(file_out)) with open(file_out, 'wb') as save_file: save_file.write(rows_to_csv(rows).read()) if 'storage' in destination and destination['storage'].get( 'bucket') and destination['storage'].get('path'): # create the bucket bucket_create(auth, project.id, destination['storage']['bucket']) # put the file file_out = destination['storage']['bucket'] + ':' + destination[ 'storage']['path'] + variant + filename if project.verbose: print 'SAVING', file_out object_put(auth, file_out, rows_to_csv(rows)) # deprecated do not use if 'trix' in destination: trix_update(auth, destination['trix']['sheet_id'], destination['trix']['sheet_range'], rows_to_csv(rows), destination['trix']['clear']) if 'email' in destination: pass if 'sftp' in destination: try: sys.stderr = StringIO() cnopts = pysftp.CnOpts() cnopts.hostkeys = None file_prefix = 'report' if 'file_prefix' in destination['sftp']: file_prefix = destination['sftp'].get('file_prefix') del destination['sftp']['file_prefix'] #sftp_configs = destination['sftp'] #sftp_configs['cnopts'] = cnopts #sftp = pysftp.Connection(**sftp_configs) sftp = pysftp.Connection(host=destination['sftp']['host'], username=destination['sftp']['username'], password=destination['sftp']['password'], port=destination['sftp']['port'], cnopts=cnopts) if 'directory' in destination['sftp']: sftp.cwd(destination['sftp']['directory']) tmp_file_name = '/tmp/%s_%s.csv' % ( file_prefix, datetime.datetime.now().strftime('%Y-%m-%dT%H-%M-%S')) tmp_file = open(tmp_file_name, 'wb') tmp_file.write(rows_to_csv(rows).read()) tmp_file.close() sftp.put(tmp_file_name) os.remove(tmp_file_name) sys.stderr = sys.__stderr__ except e: print e traceback.print_exc()
def put_rows(config, auth, destination, rows, schema=None, variant=''): """Processes standard write JSON block for dynamic export of data. Allows us to quickly write the results of a script to a destination. For example write the results of a DCM report into BigQuery. - Will write to multiple destinations if specified. - Extensible, add a handler to define a new destination ( be kind update the documentation json ). Include the following JSON in a recipe, then in the run.py handler when encountering that block pass it to this function and use the returned results. from utils.data import put_rows var_json = { "out":{ "bigquery":{ "auth":"[ user or service ]", "dataset": [ string ], "table": [ string ] "schema": [ json - standard bigquery schema json ], "header": [ boolean - true if header exists in rows ] "disposition": [ string - same as BigQuery documentation ] }, "sheets":{ "auth":"[ user or service ]", "sheet":[ string - full URL, suggest using share link ], "tab":[ string ], "range":[ string - A1:A notation ] "append": [ boolean - if sheet range should be appended to ] "delete": [ boolean - if sheet range should be cleared before writing ] ] }, "storage":{ "auth":"[ user or service ]", "bucket": [ string ], "path": [ string ] }, "file":[ string - full path to place to write file ] } } values = put_rows('user', var_json) Args: auth: (string) The type of authentication to use, user or service. rows: ( iterator ) The list of rows to be written, if NULL no action is performed. schema: (json) A bigquery schema definition. destination: (json) A json block resembling var_json described above. rows ( list ) The data being written as a list object. variant (string) Appended to destination to differentieate multiple objects Returns: If unnest is False: Returns a list of row values [[v1], [v2], ... ] If unnest is True: Returns a list of values [v1, v2, ...] """ if rows is None: if config.verbose: print('PUT ROWS: Rows is None, ignoring write.') return if 'bigquery' in destination: if not schema: schema = destination['bigquery'].get('schema') skip_rows = 1 if destination['bigquery'].get('header') and schema else 0 if destination['bigquery'].get('format', 'CSV') == 'JSON': json_to_table( config, destination['bigquery'].get('auth', auth), destination['bigquery'].get('project_id', config.project), destination['bigquery']['dataset'], destination['bigquery']['table'] + variant, rows, schema, destination['bigquery'].get('disposition', 'WRITE_TRUNCATE'), ) elif destination['bigquery'].get('is_incremental_load', False) == True: incremental_rows_to_table( config, destination['bigquery'].get('auth', auth), destination['bigquery'].get('project_id', config.project), destination['bigquery']['dataset'], destination['bigquery']['table'] + variant, rows, schema, destination['bigquery'].get('skip_rows', skip_rows), destination['bigquery'].get('disposition', 'WRITE_APPEND'), billing_project_id=config.project) else: rows_to_table( config, destination['bigquery'].get('auth', auth), destination['bigquery'].get('project_id', config.project), destination['bigquery']['dataset'], destination['bigquery']['table'] + variant, rows, schema, destination['bigquery'].get('skip_rows', skip_rows), destination['bigquery'].get('disposition', 'WRITE_TRUNCATE'), ) if 'sheets' in destination: if destination['sheets'].get('delete', False): sheets_clear( config, destination['sheets'].get('auth', auth), destination['sheets']['sheet'], destination['sheets']['tab'] + variant, destination['sheets']['range'], ) sheets_write( config, destination['sheets'].get('auth', auth), destination['sheets']['sheet'], destination['sheets']['tab'] + variant, destination['sheets']['range'], rows_to_type(rows), destination['sheets'].get('append', False), ) if 'file' in destination: path_out, file_ext = destination['file'].rsplit('.', 1) file_out = path_out + variant + '.' + file_ext if config.verbose: print('SAVING', file_out) makedirs_safe(parse_path(file_out)) with open(file_out, 'w') as save_file: save_file.write(rows_to_csv(rows).read()) if 'storage' in destination and destination['storage'].get( 'bucket') and destination['storage'].get('path'): bucket_create( config, destination['storage'].get('auth', auth), config.project, destination['storage']['bucket'] ) # put the file file_out = destination['storage']['bucket'] + ':' + destination['storage'][ 'path'] + variant if config.verbose: print('SAVING', file_out) object_put(config, auth, file_out, rows_to_csv(rows)) if 'sftp' in destination: try: cnopts = pysftp.CnOpts() cnopts.hostkeys = None path_out, file_out = destination['sftp']['file'].rsplit('.', 1) file_out = path_out + variant + file_out sftp = pysftp.Connection( host=destination['sftp']['host'], username=destination['sftp']['username'], password=destination['sftp']['password'], port=destination['sftp']['port'], cnopts=cnopts) if '/' in file_out: dir_out, file_out = file_out.rsplit('/', 1) sftp.cwd(dir_out) sftp.putfo(rows_to_csv(rows), file_out) except e: print(str(e)) traceback.print_exc()