-
Notifications
You must be signed in to change notification settings - Fork 73
/
gh-data.py
152 lines (123 loc) · 4.19 KB
/
gh-data.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
#!/usr/bin/env python
# Based on https://github.com/WebKit/standards-positions/blob/main/summary.py
import json, os, requests, re, sys
# Utilities
def write_json(filename, data):
with open(filename, "w") as f:
json.dump(data, f, indent=2, separators=(",", ": "))
f.write("\n")
# General processing
def process(issues):
summary = []
for issue in issues:
if is_ignorable_issue(issue):
continue
summary_item = {"issue": int(issue["html_url"][issue["html_url"].rfind("/") + 1 :])}
summary_item.update(process_labels(issue["labels"]))
summary_item.update(process_body(issue))
summary_item["title"] = re.sub(r"(request for (mozilla )?position|rfp)( ?:| on) ", "", issue["title"], flags=re.IGNORECASE)
summary.append(summary_item)
write_json("gh-data-summary.json", summary)
print("Done: gh-data-summary.json.")
def is_ignorable_issue(issue):
if "pull_request" in issue:
return True
for label in issue["labels"]:
if label["name"] in ("duplicate", "invalid", "tooling", "proposal withdrawn"):
return True
return False
def process_labels(labels):
position = None
venues = []
concerns = []
topics = []
for label in labels:
# Position
if label["name"].startswith("position: "):
position = label["name"].split(": ")[1]
# Venue
elif label["name"].startswith("venue: "):
venues.append(label["name"].split(": ")[1])
# Concerns
elif label["name"].startswith("concerns: "):
concerns.append(label["name"].split(": ")[1])
# Topics
elif label["name"].startswith("topic: "):
topics.append(label["name"].split(": ")[1])
return {
"position": position,
"venues": venues,
"concerns": concerns,
"topics": topics,
}
def process_body(issue):
lines = issue["body"].splitlines()
body = {
"title": None,
"url": None,
"explainer": None,
"caniuse": None,
"bug": None,
"webkit": None,
}
mapping = {
# "specification title": "title", # Always use the issue title
"specification or proposal url (if available)": "url",
"specification or proposal url": "url",
"explainer url (if available)": "explainer",
"explainer url": "explainer",
"caniuse.com url (optional)": "caniuse",
"caniuse.com url": "caniuse",
"bugzilla url (optional)": "bug",
"bugzilla url": "bug",
"webkit standards-position": "webkit",
}
for line in lines:
if line == "### Other information":
break
for title, key in mapping.items():
text_title = f"* {title}: "
if line.lower().startswith(text_title):
value = line[len(text_title) :].strip()
value = re.sub(r"\[[^\]]+\]\(([^\)]+)\)", r"\1", value)
if value != "" and value.lower() != "n/a":
body[key] = value
break
return body
# Setup
def main():
# update
data = []
page = 1
while True:
try:
print(f"Fetching page {page}...")
response = requests.get(
f"https://api.github.com/repos/mozilla/standards-positions/issues?direction=asc&state=all&per_page=100&page={page}",
timeout=5,
)
response.raise_for_status()
except Exception:
print("Update failed, network failure or request timed out.")
exit(1)
temp_data = response.json()
if not temp_data:
print("Empty!")
break
data.extend(temp_data)
# Check for 'link' header and 'rel="next"'
link_header = response.headers.get("link", "")
if 'rel="next"' not in link_header:
break
page += 1
write_json("gh-data.json", data)
print("Done: gh-data.json.")
# process
if not os.path.exists("gh-data.json"):
print("Sorry, you have to update first.")
exit(1)
with open("gh-data.json", "rb") as f:
data = json.load(f)
process(data)
if __name__ == "__main__":
main()