EricR401S commited on
Commit
ec42311
1 Parent(s): 8f25a0c
Pill_Ideologies-Post_Titles.py CHANGED
@@ -164,24 +164,33 @@ class SubRedditPosts(datasets.GeneratorBasedBuilder):
164
  urls = _URLS[self.config.name]
165
  data_dir = dl_manager.download_and_extract(urls)
166
  data = pd.read_csv(data_dir)
 
 
167
 
168
- def clean_data_nans(df):
169
- """This function takes a dataframe and fills all NaNs with a value
170
- This is to appease google colab, because my local machine did not raise errors
171
- ... and it's a windows. That should tell you a lot."""
172
- for col in data.columns:
173
- print(f"Cleaning NaNs in {col}")
174
- if df[col].dtype == "object":
175
- df[col].fillna("No-NAN-Nothing found", inplace=True)
176
- elif df[col].dtype in ["int64", "float64", "int32", "float32", "int"]:
177
- df[col].fillna(0, inplace=True)
178
- elif df[col].dtype == "bool":
179
- df[col].fillna(False, inplace=True)
180
- else:
181
- df[col].fillna("NAN - problematic {col} found", inplace=True)
182
- return None
183
-
184
- clean_data_nans(data)
 
 
 
 
 
 
 
185
  print("PAssed the cleaning")
186
  # commented out the splits, due to google colab being uncooperative
187
  # raised too many errors that my local machine did not
@@ -264,3 +273,28 @@ class SubRedditPosts(datasets.GeneratorBasedBuilder):
264
  "upvote_ratio": row.get("upvote_ratio"),
265
  "is_video": row.get("is_video"),
266
  }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
164
  urls = _URLS[self.config.name]
165
  data_dir = dl_manager.download_and_extract(urls)
166
  data = pd.read_csv(data_dir)
167
+ data = self.process_data(data)
168
+ print("PAssed the processing")
169
 
170
+ # def clean_data_nans(df):
171
+ # """This function takes a dataframe and fills all NaNs with a value
172
+ # This is to appease google colab, because my local machine did not raise errors
173
+ # ... and it's a windows. That should tell you a lot."""
174
+ # for col in data.columns:
175
+ # print(f"Cleaning NaNs in {col}")
176
+ # if df[col].dtype == "object":
177
+ # df[col].fillna("No-NAN-Nothing found", inplace=True)
178
+ # elif df[col].dtype in ["int64", "float64", "int32", "float32", "int"]:
179
+ # df[col].fillna(0, inplace=True)
180
+ # elif df[col].dtype == "bool":
181
+ # df[col].fillna(False, inplace=True)
182
+ # else:
183
+ # df[col].fillna("NAN - problematic {col} found", inplace=True)
184
+ # return None
185
+
186
+ def show_nans_profile(df):
187
+
188
+ for col in df.columns:
189
+
190
+ print(col, df[col].isna().sum())
191
+
192
+ show_nans_profile(data)
193
+ # clean_data_nans(data)
194
  print("PAssed the cleaning")
195
  # commented out the splits, due to google colab being uncooperative
196
  # raised too many errors that my local machine did not
 
273
  "upvote_ratio": row.get("upvote_ratio"),
274
  "is_video": row.get("is_video"),
275
  }
276
+
277
+ def process_data(df):
278
+ """This function takes a dataframe and processes it to remove any unwanted columns and rows"""
279
+ # columns that cannot be empty,so drop rows
280
+ df = df.dropna(subset=["subreddit"])
281
+ df = df.dropna(subset=["title"])
282
+
283
+ # cleaning to make colab importing the dataset through huggingface work
284
+ values = {
285
+ "id": "",
286
+ "text": "",
287
+ "url": "",
288
+ "score": 0,
289
+ "date": 0.0,
290
+ "subreddit_subscribers": 0,
291
+ "num_comments": 0,
292
+ "ups": 0,
293
+ "downs": 0,
294
+ "upvote_ratio": 0.0,
295
+ "is_video": False,
296
+ }
297
+
298
+ df.fillna(value=values, inplace=True)
299
+
300
+ return df
reddit_posts_fm.csv CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2517556a2189140bf673ff9ef8a36a21e6e613feb21a3524b83cfbc3cd819e12
3
- size 11273908
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c1c32caf7adba027c63be0a0bc8263bad86ab30049290b53b458eb235a69d115
3
+ size 11259261
redditscraper_fm.py CHANGED
@@ -121,7 +121,6 @@ def pull_info_from_reddit_dict(
121
  "ups",
122
  "downs",
123
  "upvote_ratio",
124
- "num_reports",
125
  "is_video",
126
  ],
127
  ):
@@ -151,39 +150,15 @@ df = pd.DataFrame(
151
  "text",
152
  "url",
153
  "score",
154
- "author",
155
  "date",
156
  "subreddit_subscribers",
157
  "num_comments",
158
  "ups",
159
  "downs",
160
  "upvote_ratio",
161
- "num_reports",
162
  "is_video",
163
  ],
164
  )
165
 
166
- # columns that cannot be empty,so drop rows
167
- df = df.dropna(subset=["subreddit"])
168
- df = df.dropna(subset=["title"])
169
- df = df.drop(columns=["num_reports"]) # drop num_reports, always empty
170
-
171
- # cleaning to make colab importing the dataset through huggingface work
172
- values = {
173
- "id": "NOTEXT",
174
- "text": "NOTEXT",
175
- "url": "NOTEXT",
176
- "score": 0,
177
- "date": 0.0,
178
- "subreddit_subscribers": 0,
179
- "num_comments": 0,
180
- "ups": 0,
181
- "downs": 0,
182
- "upvote_ratio": 0.0,
183
- "is_video": "False",
184
- }
185
- df.fillna(value=values, inplace=True)
186
-
187
- df = df[df["subreddit"].isin(subreddit_list)]
188
-
189
  df.to_csv("reddit_posts_fm.csv", index=False)
 
121
  "ups",
122
  "downs",
123
  "upvote_ratio",
 
124
  "is_video",
125
  ],
126
  ):
 
150
  "text",
151
  "url",
152
  "score",
153
+ # "author", author is not useful for the analysis
154
  "date",
155
  "subreddit_subscribers",
156
  "num_comments",
157
  "ups",
158
  "downs",
159
  "upvote_ratio",
 
160
  "is_video",
161
  ],
162
  )
163
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
164
  df.to_csv("reddit_posts_fm.csv", index=False)