trifacta 2.1

Trifacta client

trifacta

Trifacta client that makes it easy to integrate Trifacta into your production and data science workflows

Usage Scenarios

• Jupyter: Invoke Trifacta jobs from a Jupyter notebook and pass data back and forth between Jupyter and Trifacta
• Other Notebooks: Integrate Trifacta with Azure Databricks, Zepellin or any other notebook-style interface that supports Python
• Scripts: Automate Trifacta jobs and input/output using python scripts that can be easily executed from the command line or called from an external scheduler

Functionality

This library makes it simple to do the following:

1. Connect to a Trifacta instance
2. Run a job
3. Download results to a pandas dataframe OR Download results as text/csv
4. Upload files to Trifacta

Note that file uploads and downloads are performed using httpfs, and require that port 14000 be opened on the Trifacta server

#!pip install trifacta
import trifacta

#Step 1: Connect to Trifacta by providing the URL, username and password
t = trifacta.Client('https://partnerdemo.trifacta.net', '[e-mail]', 'mypassword')

Get the wrangled dataset id from the URL in the Trifacta UI

Make sure that you have run the job manually at least once

Note the output path (be sure to set it to "replace")

#Step 2: Run the job
t.run_job(14478)

About to run job
{'jobgroupId': 3926, 'jobIds': [7513, 7514], 'reason': 'JobStarted', 'sessionId': 'b9d327f0-8e19-11e8-8feb-9fabf204e996'}
2018-07-22 18:43:01.427594 InProgress
2018-07-22 18:43:06.791576 Complete





True

#Step 3a: Get a pandas dataframe with the results
df = t.get_dataframe('/trifacta/queryResults/[e-mail]/demo_output.csv')

df

	Neighborhood	HouseStyle	row_count	sum_LotArea
0	NAmes	1Story	159	1589811
1	CollgCr	1Story	91	841644
2	Gilbert	2Story	60	668112
3	Timber	1Story	23	554694
4	CollgCr	2Story	53	546602
5	NridgHt	1Story	51	537687
6	Sawyer	1Story	53	528438
7	Edwards	1Story	53	511296
8	NoRidge	2Story	33	485691
9	NWAmes	1Story	35	403813
10	ClearCr	1Story	11	395797
11	Mitchel	1Story	32	394436
12	Somerst	1Story	37	350820
13	NWAmes	2Story	29	348885
14	Somerst	2Story	49	323495
15	NridgHt	2Story	26	300685
16	OldTown	2Story	32	274465
17	SawyerW	1Story	28	271008
18	OldTown	1.5Fin	33	267283
19	ClearCr	1.5Fin	6	266593
20	Crawfor	1Story	19	260639
21	SawyerW	2Story	25	255102
22	NAmes	2Story	22	249793
23	OldTown	1Story	33	240257
24	Edwards	1.5Fin	22	228970
25	Crawfor	2Story	20	222029
26	NAmes	SLvl	21	221177
27	Edwards	2Story	14	185799
28	Timber	1.5Fin	2	178418
29	BrkSide	1.5Fin	25	172233
...	...	...	...	...
66	CollgCr	SLvl	3	30135
67	BrDale	2Story	16	28816
68	Veenker	SLvl	2	25757
69	NoRidge	1.5Fin	2	25398
70	SawyerW	SFoyer	3	25267
71	CollgCr	SFoyer	3	24491
72	MeadowV	2Story	8	19611
73	NPkVill	1Story	4	17942
74	Veenker	2Story	1	17542
75	NAmes	1.5Unf	2	16827
76	SWISU	1Story	2	14692
77	OldTown	SFoyer	2	14179
78	NWAmes	1.5Fin	1	13837
79	SawyerW	SLvl	1	12800
80	IDOTRR	1.5Unf	2	12449
81	SawyerW	1.5Fin	1	12327
82	Gilbert	1.5Fin	1	12134
83	BrkSide	2.5Unf	1	11888
84	Crawfor	2.5Fin	1	11526
85	NPkVill	2Story	5	11465
86	NWAmes	SFoyer	1	10625
87	Crawfor	1.5Unf	1	10594
88	OldTown	1.5Unf	2	9888
89	MeadowV	SFoyer	6	9853
90	SawyerW	1.5Unf	1	9000
91	MeadowV	1Story	2	8448
92	IDOTRR	2.5Unf	1	7200
93	Crawfor	2.5Unf	1	7128
94	Blueste	2Story	2	3250
95	MeadowV	SLvl	1	1596

96 rows × 4 columns

#Step 3b: Download results as text/csv
file_contents = t.get_file_contents('/trifacta/queryResults/[e-mail]/demo_output.csv')
with open('demo_output.csv', 'w') as f:
    f.write(file_contents)

#Show the first few rows of the CSV file
!head demo_output.csv

"Neighborhood","HouseStyle","row_count","sum_LotArea"
"NAmes","1Story","159","1589811"
"CollgCr","1Story","91","841644"
"Gilbert","2Story","60","668112"
"Timber","1Story","23","554694"
"CollgCr","2Story","53","546602"
"NridgHt","1Story","51","537687"
"Sawyer","1Story","53","528438"
"Edwards","1Story","53","511296"
"NoRidge","2Story","33","485691"

#Step 4: Upload files to Trifacta
t.put_file_contents('/trifacta/uploads/demo_output.csv', file_contents)

True