Skip to content

Commit 923852e

Browse files
dshemetovdshemetov
committed
style(black): format wiki acquisition
1 parent 07ed83e commit 923852e

File tree

5 files changed

+674
-606
lines changed

5 files changed

+674
-606
lines changed

‎src/acquisition/wiki/wiki.py‎

Lines changed: 121 additions & 125 deletions
Original file line numberDiff line numberDiff line change
@@ -1,112 +1,112 @@
11
"""
2-
===============
3-
=== Purpose ===
4-
===============
5-
6-
Wrapper for the entire wiki data collection process:
7-
1. Uses wiki_update.py to fetch metadata for new access logs
8-
2. Uses wiki_download.py to download the access logs
9-
3. Uses wiki_extract.py to store article access counts
10-
2+
===============
3+
=== Purpose ===
4+
===============
5+
6+
Wrapper for the entire wiki data collection process:
7+
1. Uses wiki_update.py to fetch metadata for new access logs
8+
2. Uses wiki_download.py to download the access logs
9+
3. Uses wiki_extract.py to store article access counts
10+
1111
See also: master.php
12-
13-
14-
=======================
15-
=== Data Dictionary ===
16-
=======================
17-
18-
`wiki_raw` is a staging table where extracted access log data is stored for
19-
further processing. When wiki_update.py finds a new log, it saves the name and
20-
hash to this table, with a status of 0. This table is read by master.php, which
21-
then hands out "jobs" (independently and in parallel) to wiki_download.py.
22-
After wiki_download.py downloads the log and extracts the counts, it submits
23-
the data (as JSON) to master.php, which then stores the "raw" JSON counts in
24-
this table.
25-
+----------+---------------+------+-----+---------+----------------+
26-
| Field | Type | Null | Key | Default | Extra |
27-
+----------+---------------+------+-----+---------+----------------+
28-
| id | int(11) | NO | PRI | NULL | auto_increment |
29-
| name | varchar(64) | NO | UNI | NULL | |
30-
| hash | char(32) | NO | | NULL | |
31-
| status | int(11) | NO | MUL | 0 | |
32-
| size | int(11) | YES | | NULL | |
33-
| datetime | datetime | YES | | NULL | |
34-
| worker | varchar(256) | YES | | NULL | |
35-
| elapsed | float | YES | | NULL | |
36-
| data | varchar(2048) | YES | | NULL | |
37-
+----------+---------------+------+-----+---------+----------------+
38-
id: unique identifier for each record
39-
name: name of the access log
40-
hash: md5 hash of the file, as reported by the dumps site (all zeroes if no
41-
hash is provided)
42-
status: the status of the job, using the following values:
43-
0: queued for download
44-
1: download in progress
45-
2: queued for extraction
46-
3: extracted to `wiki` table
47-
(any negative value indicates failure)
48-
size: the size, in bytes, of the downloaded file
49-
datetime: the timestamp of the most recent status update
50-
worker: name (user@hostname) of the machine working on the job
51-
elapsed: time, in seconds, taken to complete the job
52-
data: a JSON string containing counts for selected articles in the access log
53-
54-
`wiki` is the table where access counts are stored (parsed from wiki_raw). The
55-
"raw" JSON counts are parsed by wiki_extract.py and stored directly in this
56-
table.
57-
+----------+-------------+------+-----+---------+----------------+
58-
| Field | Type | Null | Key | Default | Extra |
59-
+----------+-------------+------+-----+---------+----------------+
60-
| id | int(11) | NO | PRI | NULL | auto_increment |
61-
| datetime | datetime | NO | MUL | NULL | |
62-
| article | varchar(64) | NO | MUL | NULL | |
63-
| count | int(11) | NO | | NULL | |
64-
+----------+-------------+------+-----+---------+----------------+
65-
id: unique identifier for each record
66-
datetime: UTC timestamp (rounded to the nearest hour) of article access
67-
article: name of the article
68-
count: number of times the article was accessed in the hour
69-
70-
`wiki_meta` is a metadata table for this dataset. It contains pre-calculated
71-
date and epiweeks fields, and more importantly, the total number of English
72-
article hits (denominator) for each `datetime` in the `wiki` table. This table
73-
is populated in parallel with `wiki` by the wiki_extract.py script.
74-
+----------+----------+------+-----+---------+----------------+
75-
| Field | Type | Null | Key | Default | Extra |
76-
+----------+----------+------+-----+---------+----------------+
77-
| id | int(11) | NO | PRI | NULL | auto_increment |
78-
| datetime | datetime | NO | UNI | NULL | |
79-
| date | date | NO | | NULL | |
80-
| epiweek | int(11) | NO | | NULL | |
81-
| total | int(11) | NO | | NULL | |
82-
+----------+----------+------+-----+---------+----------------+
83-
id: unique identifier for each record
84-
datetime: UTC timestamp (rounded to the nearest hour) of article access
85-
date: the date portion of `datetime`
86-
epiweek: the year and week containing `datetime`
87-
total: total number of English article hits in the hour
88-
89-
90-
=================
91-
=== Changelog ===
92-
=================
93-
12+
13+
14+
=======================
15+
=== Data Dictionary ===
16+
=======================
17+
18+
`wiki_raw` is a staging table where extracted access log data is stored for
19+
further processing. When wiki_update.py finds a new log, it saves the name and
20+
hash to this table, with a status of 0. This table is read by master.php, which
21+
then hands out "jobs" (independently and in parallel) to wiki_download.py.
22+
After wiki_download.py downloads the log and extracts the counts, it submits
23+
the data (as JSON) to master.php, which then stores the "raw" JSON counts in
24+
this table.
25+
+----------+---------------+------+-----+---------+----------------+
26+
| Field | Type | Null | Key | Default | Extra |
27+
+----------+---------------+------+-----+---------+----------------+
28+
| id | int(11) | NO | PRI | NULL | auto_increment |
29+
| name | varchar(64) | NO | UNI | NULL | |
30+
| hash | char(32) | NO | | NULL | |
31+
| status | int(11) | NO | MUL | 0 | |
32+
| size | int(11) | YES | | NULL | |
33+
| datetime | datetime | YES | | NULL | |
34+
| worker | varchar(256) | YES | | NULL | |
35+
| elapsed | float | YES | | NULL | |
36+
| data | varchar(2048) | YES | | NULL | |
37+
+----------+---------------+------+-----+---------+----------------+
38+
id: unique identifier for each record
39+
name: name of the access log
40+
hash: md5 hash of the file, as reported by the dumps site (all zeroes if no
41+
hash is provided)
42+
status: the status of the job, using the following values:
43+
0: queued for download
44+
1: download in progress
45+
2: queued for extraction
46+
3: extracted to `wiki` table
47+
(any negative value indicates failure)
48+
size: the size, in bytes, of the downloaded file
49+
datetime: the timestamp of the most recent status update
50+
worker: name (user@hostname) of the machine working on the job
51+
elapsed: time, in seconds, taken to complete the job
52+
data: a JSON string containing counts for selected articles in the access log
53+
54+
`wiki` is the table where access counts are stored (parsed from wiki_raw). The
55+
"raw" JSON counts are parsed by wiki_extract.py and stored directly in this
56+
table.
57+
+----------+-------------+------+-----+---------+----------------+
58+
| Field | Type | Null | Key | Default | Extra |
59+
+----------+-------------+------+-----+---------+----------------+
60+
| id | int(11) | NO | PRI | NULL | auto_increment |
61+
| datetime | datetime | NO | MUL | NULL | |
62+
| article | varchar(64) | NO | MUL | NULL | |
63+
| count | int(11) | NO | | NULL | |
64+
+----------+-------------+------+-----+---------+----------------+
65+
id: unique identifier for each record
66+
datetime: UTC timestamp (rounded to the nearest hour) of article access
67+
article: name of the article
68+
count: number of times the article was accessed in the hour
69+
70+
`wiki_meta` is a metadata table for this dataset. It contains pre-calculated
71+
date and epiweeks fields, and more importantly, the total number of English
72+
article hits (denominator) for each `datetime` in the `wiki` table. This table
73+
is populated in parallel with `wiki` by the wiki_extract.py script.
74+
+----------+----------+------+-----+---------+----------------+
75+
| Field | Type | Null | Key | Default | Extra |
76+
+----------+----------+------+-----+---------+----------------+
77+
| id | int(11) | NO | PRI | NULL | auto_increment |
78+
| datetime | datetime | NO | UNI | NULL | |
79+
| date | date | NO | | NULL | |
80+
| epiweek | int(11) | NO | | NULL | |
81+
| total | int(11) | NO | | NULL | |
82+
+----------+----------+------+-----+---------+----------------+
83+
id: unique identifier for each record
84+
datetime: UTC timestamp (rounded to the nearest hour) of article access
85+
date: the date portion of `datetime`
86+
epiweek: the year and week containing `datetime`
87+
total: total number of English article hits in the hour
88+
89+
90+
=================
91+
=== Changelog ===
92+
=================
93+
9494
2017-02-24
9595
* secrets and small improvements
9696
2016-08-14
9797
* Increased job limit (6 -> 12) (pageviews files are ~2x smaller)
98-
2015-08-26
98+
2015-08-26
9999
* Reduced job limit (8 -> 6)
100-
2015-08-14
100+
2015-08-14
101101
* Reduced job limit (10 -> 8)
102-
2015-08-11
102+
2015-08-11
103103
+ New table `wiki_meta`
104-
2015-05-22
104+
2015-05-22
105105
* Updated status codes for `wiki_raw` table
106-
2015-05-21
106+
2015-05-21
107107
* Original version
108108
"""
109-
109+
110110
# first party
111111
from . import wiki_update
112112
from . import wiki_download
@@ -115,31 +115,27 @@
115115

116116

117117
def main():
118-
# step 1: find new access logs (aka "jobs")
119-
print('looking for new jobs...')
120-
try:
121-
wiki_update.run()
122-
except:
123-
print('wiki_update failed')
124-
125-
# step 2: run a few jobs
126-
print('running jobs...')
127-
try:
128-
wiki_download.run(
129-
secrets.wiki.hmac,
130-
download_limit=1024 * 1024 * 1024,
131-
job_limit=12
132-
)
133-
except:
134-
print('wiki_download failed')
135-
136-
# step 3: extract counts from the staging data
137-
print('extracting counts...')
138-
try:
139-
wiki_extract.run(job_limit=100)
140-
except:
141-
print('wiki_extract failed')
142-
143-
144-
if __name__ == '__main__':
145-
main()
118+
# step 1: find new access logs (aka "jobs")
119+
print("looking for new jobs...")
120+
try:
121+
wiki_update.run()
122+
except:
123+
print("wiki_update failed")
124+
125+
# step 2: run a few jobs
126+
print("running jobs...")
127+
try:
128+
wiki_download.run(secrets.wiki.hmac, download_limit=1024 * 1024 * 1024, job_limit=12)
129+
except:
130+
print("wiki_download failed")
131+
132+
# step 3: extract counts from the staging data
133+
print("extracting counts...")
134+
try:
135+
wiki_extract.run(job_limit=100)
136+
except:
137+
print("wiki_extract failed")
138+
139+
140+
if __name__ == "__main__":
141+
main()

0 commit comments

Comments
 (0)