taskstatistic.py

from tqdm import tqdm
import json
f = open("/home/aiscuser/fhw/data/all_instruct_with_tasks.json", "r+")
lines = f.readlines()
a , b, c, d, e, f, g = 0, 0, 0, 0, 0, 0, 0
for line in tqdm(lines):
    dd = json.loads(line)
    text = dd["task"]
    if "Code Transpile" in text:
        g+=1
        continue
    if "Code Debugging" in text:
        b +=1
        continue
    if "Code Optimization" in text:
        c +=1
        continue
    if "Code Reasoning" in text:
        d +=1
        continue
    if "Code Analysis" in text:
        e +=1
        continue
    if "Code Generation" in text:
        a +=1
        continue
    if "Theoretical Explanation" in text:
        f +=1
        continue

print({"Code Generation": a, "Code Debugging": b, "Code Optimization": c, "Code Reasoning": d, "Code Analysis": e, "Theoretical Explanation": f, "Code Transpile": g})