import requests, json, re

url = "https://hexbear.net/api/v3/comment/list?post_id=2207213&sort=New&limit=50&page="

headers = {"accept": "application/json"}

yes = 0
nos = 0
maybes = 0
comment_num = 0

for i in range(1, 10):
    response = requests.get(url+str(i), headers=headers)
    comments = json.loads(response.text)["comments"]
    
    for comment in comments:
        if comment["comment"]["removed"] == False and comment["comment"]["deleted"] == False:
            comment_num += 1
            comment_content = comment["comment"]["content"]
            if re.search('^(\n?

• included: only where the first word was "yes","no","maybe"
• had intended to be case insensitive but realized after everything was done that I accidentally excluded 2 people who answered in all caps
• below in method, this is fixed so if you follow it you will get slight variation
• included: first listed pronoun only (otherwise too complicated)
  • eg: [he/him, comrade/them] = he, [she/her, they/them] = she
  • in Methods spoiler it includes a way to get all 4 if you want but it would need to be cleaned up
• excluded: text in spoilers (too complicated)
• excluded: accounts not on hexbear.net (no pronouns)
• excluded: deleted replies

• opened the thread and scrolled to the end, I think maybe selecting "chat" button at the top lets you get more responses? Also I tried sorting a couple different ways and copying everything as later it is deduplicated
• note: am using firefox on linux with extensions that might affect the results; I didn't think to disable or anything at the beginning but I would if I were to repeat
• selected all responses and copy to clipboard
• go to https://regex101.com/
• paste in the Test string text box
• on left sidebar select Function > List
• in the Regular expression field (above Test string), paste: /^(\w+) \[(\w+).*\](.*\n.*\n+)(no|yes|maybe)/gim
• in the List field, (below Test string), paste $1;$2;$4;\n
• here I used ; (semi colon) as column delimiter (used semicolon instead of , (comma) for compatibility when scraping all 4 pronouns)

expected output is like this:

useruseruser1;he;Yes;
useruseruser2;he;No;
useruseruser3;he;no;
useruseruser4;she;Yes;
useruseruser5;he;No;

• copy and paste the contents of List text box into a text file results.csv and save

Set up spreadsheet

• use a regular text editor (like notepad.exe on windows) to create a plain text file called template.csv
• paste the following:

"number","UID","pron","ans","concat","pron count","pron count %","ans count","ans count %","cat count","cat count %","keep","n"
"=RANDBETWEEN.NV(10000,99999)",,,,"=CONCAT(C2,""-"",D2)","=COUNTIF(C:C,C2)","=ROUND(F2/$M2,3)*100","=COUNTIF(D:D,D2)","=ROUND(H2/$M2,3)*100","=COUNTIF(E:E,E2)","=ROUND(J2/$M2,3)*100",1,"=COUNTIF(L:L,1)"

• Launch LibreOffice Calc (open source Excel)
• File > Open and select template.csv It should open the text import dialogue. Important settings:
  • Separated by: comma (uncheck the others)
  • String delimiter: " (double quotation marks)
  • Other options: CHECK Evaluate formulas
• This should give you a 2 row spreadsheet
• You might want to use File > Save As to save it as ODS format which allows more features compared to csv. I name this file combined.ods.

Open results

• File > Open and select results.csv. It should open the text import dialogue. Important settings:
  • Separated by: semicolon (uncheck the others) note this is different than the previous step!!
• You should get a new spreadsheet document with 3 colulmns:
  • Column A: usernames
  • Column B: Pronouns
  • Column C: Response (Yes/No/Maybe)
• Drag to select all the text
• Copy to clipboard
• Close this file (once the next part works)

Put results in template

• Go back to your other file, combined.ods
• Select all content of row 2
• drag to fill down several hundred rows (however many results you have)
• Place cursor in cell B2 (blank under column heading UID
• Paste clipboard
• You should get 300something rows inserted to columns B, C, D
• You must delete the extra template rows (otherwise your n will be wrong)
• You must manually edit the column called n (the last column, actually it is column number M) so that each value is 1. by default it will start to count so the values will because 1, 2, 3 etc. Should be 1,1,1.
  • This column controls whether the row is included in the data. If you set the value for 0 it will be excluded. So you can keep data you want to exclude in the sheet.

Now you must know how to use a spreadsheet. I suggest fixing all the answers to be the same case otherwise Yes and yes might be different in some circumstances.

Briefly the columns are: -A: number: assigns a random number so the file can be shared without publishing usernames - anonymizes data but lets you keep a record of how you got the results. How to:

• open a copy of file
• delete the column UID
• sort by column number (otherwise results will still be in the order as on the page and it might be possible to infer information)
• I didn't include it in the template but I actually also checked there were not any duplicate numbers also; you should do this
• B: UID: user name
• C: pron: first pronoun
• D: ans: the answer provided -E: concat: combines columns C and D. Expected output like she-Yes. This is respresented in "Pronoun answer combinations" -F: pron count, H: ans count, J: cat count: absolute number of each -G: pron count %, I: ans count %, K: cat count %: the count value of each as a % of *n* -L: keep: whether to include this line (explained elsewhere) -M: n`: total number of included rows

• in the Regular expression field (above Test string), paste: /^(\w+) \[(.*)\](.*\n.*\n+)(no|yes|maybe|deleted)/gim

expected output is like this:

useruseruser1;he/him, they/them;Yes;
useruseruser2;he/him;No;
useruseruser3;he/him, des/pair;no;
useruseruse4r;she/her;Yes;
useruseruser5;he/him;No;

subsequent instructions such as the spreadsheet may need adjustment