-
Notifications
You must be signed in to change notification settings - Fork 0
/
2.vectors.py
191 lines (146 loc) · 5.41 KB
/
2.vectors.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
#!/usr/bin/env python
from gensim.models.doc2vec import Doc2Vec, TaggedDocument
from scipy.spatial.distance import pdist, squareform
from sklearn import manifold
from glob import glob
import pandas
import umap
import argparse
import sys
import os
sys.path.insert(0, os.getcwd())
from helpers import process_text, write_json, read_json, read_errors
def get_parser():
parser = argparse.ArgumentParser(
description="Spack Monitor Analyser",
formatter_class=argparse.RawTextHelpFormatter,
)
parser.add_argument(
"--data_dir",
help="Directory with data",
default=os.path.join(os.getcwd(), "data"),
)
return parser
def build_model(texts, name, outdir):
# 40 epochs means we do it 40 times
model = Doc2Vec(texts, vector_size=50, min_count=5, workers=4, epochs=40)
# Save the model if we need again
model.save(os.path.join(outdir, "model.%s.doc2vec" % name))
# Create a vector for each document
# UIDS as id for each row, vectors across columns
df = pandas.DataFrame(columns=range(50))
print("Generating vector matrix for documents...")
for text in texts:
df.loc[text.tags[0]] = model.infer_vector(text.words)
# Save dataframe to file
df.to_csv(os.path.join(outdir, "%s-vectors.csv" % name))
# Create a distance matrix
distance = pandas.DataFrame(
squareform(pdist(df)), index=list(df.index), columns=list(df.index)
)
distance.to_csv(os.path.join(outdir, "%s-software-distances.csv" % name))
# Try umap first...
reducer = umap.UMAP()
embedding = reducer.fit_transform(distance)
emb = pandas.DataFrame(embedding, index=distance.index, columns=["x", "y"])
emb.index.name = "name"
emb.to_csv(os.path.join("docs", "%s-umap-software-embeddings.csv" % name))
# Make the tsne (output embeddings go into docs for visual)
fit = manifold.TSNE(n_components=2)
embedding = fit.fit_transform(distance)
emb = pandas.DataFrame(embedding, index=distance.index, columns=["x", "y"])
emb.index.name = "name"
emb.to_csv(os.path.join("docs", "%s-software-embeddings.csv" % name))
def main():
parser = get_parser()
args, extra = parser.parse_known_args()
# Make sure output directory exists
datadir = os.path.abspath(args.data_dir)
if not os.path.exists(datadir):
sys.exit("%s does not exist!" % datadir)
# Build model with errors
errors = read_errors(datadir)
# Make model output directory
model_dir = os.path.join(datadir, "models")
if not os.path.exists(model_dir):
os.makedirs(model_dir)
if not os.path.exists("docs"):
os.makedirs("docs")
# Generate metadata for errors and warnings (lookup of ID)
meta = {}
for entry in errors:
# Split based on error, add as metadata
# Add error and error parsed if relevant (we only do this for text)
if entry.get("text") and "error:" in entry["text"]:
entry["error"] = entry["text"].split("error:")[-1]
text = entry.get("text").split("error:", 1)[-1]
entry["error_parsed"] = " ".join(process_text(text))
# Add all three parsed
entry["pre_parsed"] = " ".join(process_text(entry.get("pre_context")))
entry["post_parsed"] = " ".join(process_text(entry.get("post_context")))
entry["text_parsed"] = " ".join(process_text(entry.get("text")))
meta[entry["id"]] = entry
write_json(meta, os.path.join("docs", "meta.json"))
# Try JUST the error message (e.g., error:)
texts = []
for entry in errors:
# Pre, text, and post
text = entry.get("text")
if not text:
continue
# Split based on error
if "error:" not in text:
continue
text = text.split("error:", 1)[-1]
tokens = process_text(text)
texts.append(TaggedDocument(tokens, [entry["id"]]))
build_model(texts, "error", model_dir)
# Let's try creating three models: first pre, text, and post
texts = []
for entry in errors:
# Pre, text, and post
text = (
entry.get("pre_context", "")
+ " "
+ entry.get("text")
+ " "
+ entry.get("post_context")
)
if not text:
continue
tokens = process_text(text)
meta[entry["id"]] = entry
texts.append(TaggedDocument(tokens, [entry["id"]]))
build_model(texts, "pre-text-post", model_dir)
# pre and text
texts = []
for entry in errors:
# Pre, text
text = entry.get("pre_context", "") + " " + entry.get("text")
if not text:
continue
tokens = process_text(text)
texts.append(TaggedDocument(tokens, [entry["id"]]))
build_model(texts, "pre-text", model_dir)
# post and text
texts = []
for entry in errors:
# Pre, text
text = entry.get("text") + " " + entry.get("post_context", "")
if not text:
continue
tokens = process_text(text)
texts.append(TaggedDocument(tokens, [entry["id"]]))
build_model(texts, "text-post", model_dir)
# now just text!
texts = []
for entry in errors:
# Pre, text, and post
text = entry.get("text")
if not text:
continue
tokens = process_text(text)
texts.append(TaggedDocument(tokens, [entry["id"]]))
build_model(texts, "text", model_dir)
if __name__ == "__main__":
main()