-
Notifications
You must be signed in to change notification settings - Fork 1
/
updater.py
453 lines (368 loc) · 15.9 KB
/
updater.py
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
# IMPORTS -------------------------------------------------------------------- #
import pandas as pd
import numpy as np
import requests
import json
import re
from datetime import datetime
import time
from tqdm import tqdm
from bs4 import BeautifulSoup as bs4
import warnings
warnings.simplefilter(action="ignore", category=FutureWarning)
# CONSTANTS ------------------------------------------------------------------ #
# Set constants for data provider and data API.
PROVIDER = "opendata.swiss"
PROVIDER_LINK = "https://opendata.swiss/"
BASELINK_DATAPORTAL = "https://opendata.swiss/de/dataset/"
CKAN_API_LINK = (
"https://ckan.opendata.swiss/api/3/action/current_package_list_with_resources"
)
LANGUAGE = "de"
# Set constants in regard to GitHub account and repo.
GITHUB_ACCOUNT = "rnckp"
REPO_NAME = "starter-code_opendataswiss"
REPO_BRANCH = "main"
REPO_RMARKDOWN_OUTPUT = "01_r-markdown/"
REPO_PYTHON_OUTPUT = "02_python/"
TEMP_PREFIX = "_work/"
# Set local folders and file names.
TEMPLATE_FOLDER = "_templates/"
# Template for the README.md in the repo.
TEMPLATE_README = "template_md_readme.md"
# Header for list overview that is rendered as a GitHub page.
TEMPLATE_HEADER = "template_md_header.md"
TEMPLATE_PYTHON = "template_python.ipynb"
TEMPLATE_RMARKDOWN = "template_rmarkdown.Rmd"
METADATA_FOLDER = "_metadata_json/"
TODAY_DATE = datetime.today().strftime("%Y-%m-%d")
TODAY_DATETIME = datetime.today().strftime("%Y-%m-%d %H:%M:%S")
# Set max length of dataset title in markdown table.
TITLE_MAX_CHARS = 200
# Sort markdown table by this feature.
SORT_TABLE_BY = f"title.{LANGUAGE}"
# Select keys in metadata for dataset and distributions.
KEYS_DATASET = [
"publisher",
f"organization.display_name.{LANGUAGE}",
"organization.url",
"maintainer",
"maintainer_email",
f"keywords.{LANGUAGE}",
"issued",
"metadata_created",
"metadata_modified",
]
KEYS_DISTRIBUTIONS = ["package_id", "description", "issued", "modified", "rights"]
# Select relevant column names to reduce dataset.
REDUCED_FEATURESET = [
"maintainer",
"issued",
"title_for_slug",
"maintainer_email",
"contact_points",
"id",
"metadata_created",
"metadata_modified",
"resources",
"groups",
"publisher",
"name",
"language",
"modified",
"url",
"identifier",
"keywords.fr",
"keywords.de",
"keywords.en",
"keywords.it",
"display_name.fr",
"display_name.de",
"display_name.en",
"display_name.it",
"description.fr",
"description.de",
"description.en",
"description.it",
"organization.display_name.fr",
"organization.display_name.de",
"organization.display_name.en",
"organization.display_name.it",
"organization.name",
"organization.title.fr",
"organization.title.de",
"organization.title.en",
"organization.title.it",
"title.fr",
"title.de",
"title.en",
"title.it",
"metadata",
"contact",
"distributions",
"distribution_links",
]
# FUNCTIONS ------------------------------------------------------------------ #
def get_full_package_list(limit=500, sleep=2):
"""Get full package list from CKAN API"""
offset = 0
frames = []
while True:
print(f"{offset} packages retrieved.")
url = CKAN_API_LINK + f"?limit={limit}&offset={offset}"
res = requests.get(url)
data = json.loads(res.content)
if data["result"] == []:
break
data = pd.DataFrame(pd.json_normalize(data["result"]))
frames.append(data)
offset += limit
time.sleep(sleep)
data = pd.concat(frames)
data.reset_index(drop=True, inplace=True)
return data
def has_csv_distribution(dists):
"""Iterate over package resources and keep only CSV entries in list"""
csv_dists = [x for x in dists if x.get("format", "") == "CSV"]
if csv_dists != []:
return csv_dists
else:
return np.nan
def filter_csv(data):
"""Remove all datasets that have no CSV distribution"""
data.resources = data.resources.apply(has_csv_distribution)
data.dropna(subset=["resources"], inplace=True)
return data.reset_index(drop=True)
def clean_features(data):
"""Clean various features"""
# Reduce publisher data to name-
data.publisher = data.publisher.apply(lambda x: json.loads(x)["name"])
# Reduce tags to tag names.
data.tags = data.tags.apply(lambda x: [tag["name"] for tag in x])
# Replace empty urls with NA message.
data[data["organization.url"] == ""]["organization.url"] = "None provided"
# If title in target language does not exist try to fill in english title.
idx = data[data[f"title.{LANGUAGE}"] == ""].index
data.loc[idx, f"title.{LANGUAGE}"] = data["title.en"]
# If titles are still empty try to fill in french.
idx = data[data[f"title.{LANGUAGE}"] == ""].index
data.loc[idx, f"title.{LANGUAGE}"] = data["title.fr"]
# Remove HTML tags from description.
data[f"description.{LANGUAGE}"] = data[f"description.{LANGUAGE}"].apply(
lambda x: bs4(x, "html.parser").text
)
# Strip whitespace from title.
data[f"title.{LANGUAGE}"] = data[f"title.{LANGUAGE}"].map(lambda x: x.strip())
return data
def prepare_data_for_codebooks(data):
"""Prepare metadata from catalogue in order to create the code files"""
# Add new features to save prepared data.
data["metadata"] = None
data["contact"] = None
data["distributions"] = None
data["distribution_links"] = None
# Iterate over datasets and create additional data for markdown and code cells.
for idx in tqdm(data.index):
md = [f"- **{k.capitalize()}** `{data.loc[idx, k]}`\n" for k in KEYS_DATASET]
data.loc[idx, "metadata"] = "".join(md)
if data.loc[idx, "contact_points"] != []:
contact_points = [
x for x in data.loc[idx, "contact_points"][0].values() if x != {}
]
data.loc[idx, "contact"] = " | ".join(contact_points)
else:
data.loc[idx, "contact"] = "No contact information provided."
tmp_dists = []
tmp_links = []
for dist in data.loc[idx, "resources"]:
# Some descriptions are strings rather than dicts with language keys.
if isinstance(dist["description"], dict):
# Filter description in specific language set by LANGUAGE.
if dist["description"][LANGUAGE] is not None:
# Remove line breaks of description since these break the comment blocks.
dist["description"] = re.sub(
r"[\n\r]+", " ", dist["description"][LANGUAGE]
)
else:
dist["description"] = ""
# Get other metadata of distribution.
md = [
f"# {k.capitalize():<25}: {dist.get(k, None)}\n"
for k in KEYS_DISTRIBUTIONS
]
tmp_dists.append("".join(md))
# In a few cases the dataset has no download_url but rather is available at "url".
csv_url = dist.get("download_url", dist["url"])
tmp_links.append(csv_url)
# Use .at[] – https://stackoverflow.com/a/53299945/7117003
data.at[idx, "distributions"] = tmp_dists
data.at[idx, "distribution_links"] = tmp_links
# Sort values for table.
data.sort_values(f"{SORT_TABLE_BY}", inplace=True)
data.reset_index(drop=True, inplace=True)
return data[REDUCED_FEATURESET]
def create_python_notebooks(data):
"""Create Jupyter Notebooks with Python starter code"""
for idx in tqdm(data.index):
with open(f"{TEMPLATE_FOLDER}{TEMPLATE_PYTHON}") as file:
py_nb = file.read()
# Populate template with metadata.
py_nb = py_nb.replace("{{ PROVIDER }}", PROVIDER)
title = re.sub('"', "'", data.loc[idx, f"title.{LANGUAGE}"])
py_nb = py_nb.replace("{{ DATASET_TITLE }}", title)
description = data.loc[idx, f"description.{LANGUAGE}"]
description = re.sub('"', "'", description)
description = re.sub("\\\\", "|", description)
py_nb = py_nb.replace("{{ DATASET_DESCRIPTION }}", description)
py_nb = py_nb.replace("{{ DATASET_IDENTIFIER }}", data.loc[idx, "identifier"])
py_nb = py_nb.replace(
"{{ DATASET_METADATA }}", re.sub('"', "'", data.loc[idx, "metadata"])
)
py_nb = py_nb.replace(
"{{ DISTRIBUTION_COUNT }}", str(len(data.loc[idx, "distributions"]))
)
url = f'[Direct link by {PROVIDER} for dataset]({BASELINK_DATAPORTAL}{data.loc[idx, "name"]})'
py_nb = py_nb.replace("{{ DATASHOP_LINK_PROVIDER }}", url)
if data.loc[idx, "url"] is not None:
org_name = f"organization.display_name.{LANGUAGE}"
url = data.loc[idx, "url"]
url = f"[Direct link by {data.loc[idx, org_name]} for dataset]({url})"
py_nb = py_nb.replace("{{ DATASHOP_LINK_ORGANIZATION }}", url)
py_nb = py_nb.replace("{{ CONTACT }}", data.loc[idx, "contact"])
py_nb = json.loads(py_nb, strict=False)
# Find code cell for dataset imports.
for id_cell, cell in enumerate(py_nb["cells"]):
if cell["id"] == "0":
dist_cell_idx = id_cell
break
# Iterate over csv distributions and create metadata comments and code.
code_block = []
for id_dist, (dist, dist_link) in enumerate(
zip(data.loc[idx, "distributions"], data.loc[idx, "distribution_links"])
):
code = (
f"# Distribution {id_dist}\n{dist}\ndf = get_dataset('{dist_link}')\n"
)
code = "".join([f"{line}\n" for line in code.split("\n")])
code_block.append(code)
# Populate code block with data for all distributions.
code_block = "".join(code_block)
py_nb["cells"][dist_cell_idx]["source"] = code_block
# Save to disk.
with open(
f'{TEMP_PREFIX}{REPO_PYTHON_OUTPUT}{data.loc[idx, "id"]}.ipynb',
"w",
encoding="utf-8",
) as file:
file.write(json.dumps(py_nb))
def create_rmarkdown(data):
"""Create R Markdown files with R starter code"""
for idx in tqdm(data.index):
with open(f"{TEMPLATE_FOLDER}{TEMPLATE_RMARKDOWN}") as file:
rmd = file.read()
# Populate template with metadata.
title = f"Open Government Data, {PROVIDER}"
rmd = rmd.replace("{{ DOCUMENT_TITLE }}", title)
title = re.sub('"', "'", data.loc[idx, f"title.{LANGUAGE}"])
rmd = rmd.replace("{{ DATASET_TITLE }}", title)
rmd = rmd.replace("{{ TODAY_DATE }}", TODAY_DATE)
rmd = rmd.replace("{{ DATASET_IDENTIFIER }}", data.loc[idx, "identifier"])
description = data.loc[idx, f"description.{LANGUAGE}"]
description = re.sub('"', "'", description)
description = re.sub("\\\\", "|", description)
rmd = rmd.replace("{{ DATASET_DESCRIPTION }}", description)
rmd = rmd.replace("{{ DATASET_METADATA }}", data.loc[idx, "metadata"])
rmd = rmd.replace("{{ CONTACT }}", data.loc[idx, "contact"])
rmd = rmd.replace(
"{{ DISTRIBUTION_COUNT }}", str(len(data.loc[idx, "distributions"]))
)
url = f'[Direct link by **{PROVIDER}** for dataset]({BASELINK_DATAPORTAL}{data.loc[idx, "name"]})'
rmd = rmd.replace("{{ DATASHOP_LINK_PROVIDER }}", url)
if data.loc[idx, "url"] is not None:
org_name = f"organization.display_name.{LANGUAGE}"
url = data.loc[idx, "url"]
url = f"[Direct link by **{data.loc[idx, org_name]}** for dataset]({url})"
rmd = rmd.replace("{{ DATASHOP_LINK_ORGANIZATION }}", url)
# Create code blocks for all distributions.
code_block = []
for id_dist, (dist, dist_link) in enumerate(
zip(data.loc[idx, "distributions"], data.loc[idx, "distribution_links"])
):
code = (
f"# Distribution {id_dist}\n{dist}\ndf <- read_delim('{dist_link}')\n\n"
)
code_block.append(code)
rmd = rmd.replace("{{ DISTRIBUTIONS }}", "".join(code_block))
# Save to disk.
with open(
f'{TEMP_PREFIX}{REPO_RMARKDOWN_OUTPUT}{data.loc[idx, "id"]}.Rmd',
"w",
encoding="utf-8",
) as file:
file.write("".join(rmd))
def get_header(dataset_count):
"""Retrieve header template and populate with date and count of data records"""
with open(f"{TEMPLATE_FOLDER}{TEMPLATE_HEADER}", encoding="utf-8") as file:
header = file.read()
gh_page = f"https://{GITHUB_ACCOUNT}.github.io/{REPO_NAME}/"
header = re.sub("{{ GITHUB_PAGE }}", gh_page, header)
gh_link = f"https://www.github.com/{GITHUB_ACCOUNT}/{REPO_NAME}"
header = re.sub("{{ GITHUB_REPO }}", gh_link, header)
header = re.sub("{{ PROVIDER }}", PROVIDER, header)
header = re.sub("{{ DATA_PORTAL }}", PROVIDER_LINK, header)
header = re.sub("{{ DATASET_COUNT }}", str(int(dataset_count)), header)
header = re.sub("{{ TODAY_DATE }}", TODAY_DATETIME, header)
return header
def create_readme(dataset_count):
"""Retrieve README template and populate with metadata"""
with open(f"{TEMPLATE_FOLDER}{TEMPLATE_README}", encoding="utf-8") as file:
readme = file.read()
readme = re.sub("{{ PROVIDER }}", PROVIDER, readme)
readme = re.sub("{{ DATASET_COUNT }}", str(int(dataset_count)), readme)
readme = re.sub("{{ DATA_PORTAL }}", PROVIDER_LINK, readme)
gh_page = f"https://{GITHUB_ACCOUNT}.github.io/{REPO_NAME}/"
readme = re.sub("{{ GITHUB_PAGE }}", gh_page, readme)
readme = re.sub("{{ TODAY_DATE }}", TODAY_DATETIME, readme)
with open(f"{TEMP_PREFIX}README.md", "w", encoding="utf-8") as file:
file.write(readme)
def create_overview(data, header):
"""Create README with link table"""
baselink_r_gh = f"https://github.com/{GITHUB_ACCOUNT}/{REPO_NAME}/blob/{REPO_BRANCH}/{REPO_RMARKDOWN_OUTPUT}/"
baselink_py_gh = f"https://github.com/{GITHUB_ACCOUNT}/{REPO_NAME}/blob/{REPO_BRANCH}/{REPO_PYTHON_OUTPUT}/"
baselink_py_colab = f"https://githubtocolab.com/{GITHUB_ACCOUNT}/{REPO_NAME}/blob/{REPO_BRANCH}/{REPO_PYTHON_OUTPUT}/"
# baselink_py_kaggle = f"https://kaggle.com/kernels/welcome?src={baselink_py_gh}"
md_doc = []
md_doc.append(header)
md_doc.append(
f"| Title (abbreviated to {TITLE_MAX_CHARS} chars) | Python Colab | Python GitHub | R GitHub |\n"
)
md_doc.append("| :-- | :-- | :-- | :-- |\n")
for idx in tqdm(data.index):
# Remove square brackets from title, since these break markdown links.
title_clean = (
data.loc[idx, f"title.{LANGUAGE}"].replace("[", " ").replace("]", " ")
)
if len(title_clean) > TITLE_MAX_CHARS:
title_clean = title_clean[:TITLE_MAX_CHARS] + "…"
ds_link = f'{BASELINK_DATAPORTAL}{data.loc[idx, "name"]}'
filename = data.loc[idx, "id"]
r_gh_link = f"[R GitHub]({baselink_r_gh}{filename}.Rmd)"
py_gh_link = f"[Python GitHub]({baselink_py_gh}{filename}.ipynb)"
py_colab_link = f"[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)]({baselink_py_colab}{filename}.ipynb)"
# py_kaggle_link = f'[![Kaggle](https://kaggle.com/static/images/open-in-kaggle.svg)]({baselink_py_kaggle}{filename}.ipnyb)'
md_doc.append(
f"| [{title_clean}]({ds_link}) | {py_colab_link} | {py_gh_link} | {r_gh_link} |\n"
)
md_doc = "".join(md_doc)
with open(f"{TEMP_PREFIX}index.md", "w", encoding="utf-8") as file:
file.write(md_doc)
# CREATE CODE FILES ---------------------------------------------------------- #
all_packages = get_full_package_list()
df = filter_csv(all_packages)
df = clean_features(df)
df = prepare_data_for_codebooks(df)
create_python_notebooks(df)
create_rmarkdown(df)
header = get_header(dataset_count=len(df))
create_readme(dataset_count=len(df))
create_overview(df, header)