autograders/mongodb/source/autograder.py

import pymongo, json, re
from bson import json_util
print('ver: 1.5')
# dbprep
fsroot = '/autograder/source/'
datasets = ['movies', 'comments']
db = pymongo.MongoClient('mongodb://127.0.0.1')['test']
points = (4,4,5,5,8,8,8,8)

def postproc_str(data : str): # relaxed str matching
    return re.sub(r'[\s|_|.]', '', data.lower())

def postproc_keys(data : str): 
    if type(data) is not str: 
        return postproc_iter(data)
    _dict = {}
    data = postproc_str(data)
    for k, v in _dict.items():
        data = re.sub(k, v, data)
    return data

def comparator(a, b):
    cmp = lambda x, y: 1 if x < y else -1 if x > y else 0
    try: 
        return cmp(a, b)
    except Exception:
        from collections.abc import Iterable
        def itcmp(a: Iterable, b: Iterable):
            if len(a) < len(b):
                    return -1
            elif len(a) == len(b):
                for aa, bb in zip(a, b):
                    c = comparator(aa, bb)
                    if  c != 0: 
                        return c
            else: return 1
            return 0
        from bson import ObjectId
        
        match (a, b):
            case (dict(), dict()):
                return itcmp([*a.keys(), *a.values()], [*b.keys(), *b.values()])
            case (ObjectId(aa), ObjectId(bb)):
                return cmp(aa, bb)
            case (Iterable(), Iterable()):
                return itcmp(a, b)
            case _ if type(a) == type(b):
                return cmp(f'{a}', f'{b}')
            case _:
                return cmp(hash(type(a)), hash(type(b)))

def postproc_iter(data, naive = False): 
    from collections.abc import Iterable
    from functools import cmp_to_key
    try:
        match data:
            case str():
                return postproc_str(data)
            case dict():
                return { 
                    (i if naive else postproc_keys(k)) : postproc_iter(v, naive)
                         for i, (k, v) in enumerate(data.items()) 
                }
            case Iterable(): # flatten, remove order and empty iterables
                res = type(data)(
                    sorted(
                        [postproc_iter(d, naive) for d in data 
                            if not isinstance(d, Iterable) or d]
                    , key = cmp_to_key(comparator))
                )
                return res[0] if len(res) == 1 else res
            case _: # primitives
                return data 
    except Exception as e: # fail proof
        print(e)
        return data
    


def evaluate(query : str):
    if type(query) is not str: 
        return tuple(evaluate(q) for q in query)
    query = re.sub(r'//[^\n]*', '', query)
    query = re.sub(r'(\$?[\d\w_]+)[\s\r\n]*:', r'"\1" :', query)
    query = re.sub(r'[\r\n]|.\s*pretty\s*\(\s*\)|.\s*sort\s*\([^\)]*\)', '', query).strip()
    if not query: return [None] * 2
    query = re.sub(r'.\s*aggregate\s*\(\s*([^\[^\s][^\)]*)\)', r'.aggregate([\1])', query)

    if query.endswith(';'): query = query[:-1]
    true = True
    data = list(eval(query))
    return [postproc_iter(data, n) if query else None for n in (True, False)]

for d in datasets:    
    with open(fsroot + d + '.json', encoding = 'utf-8') as f:
        # ds = re.sub(r'{\s*"\$oid"\s*:\s*("\w*")\s*}', r'ObjectId(\1)', f.read())
        jsonds = '[' + f.read().strip().replace('\n', ',') + ']'
        db[d].insert_many(json.loads(jsonds, object_hook=json_util.object_hook))

from solution import sols
answers = [evaluate(s) for s in sols]

# grading
from os import listdir
from importlib.util import module_from_spec, spec_from_file_location
subroot = '/autograder/submission/'
feedback = ''
submissions = [subroot + f for f in listdir(subroot) if f.strip().lower().endswith('.py')]

grade = 0
n_queries = len(sols)

def grade78(ans, i): 
    sol = answers[i]
    others = ('otherbill', 'otherperson')[i - 6]
    if type(ans) != list or len(ans) != len(sol):
        return False
    for a in ans:
        if a not in sol:
            if type(a) is dict:
                try:
                    for ak in a.keys():
                        if ak.startswith(others):
                            key_others = ak[len(others):]
                            t = a[key_others]
                            a[key_others] = a[ak]  
                            a[ak] = t
                    if a not in sol:
                        return False
                    else:
                        sol.remove(a)
                except Exception as e:
                    print(e)
                    return False
            else:
                return False
    return True

wa = rte = False
if submissions: 
    submission = submissions[0]
    
    for i in range(n_queries):
        feedback += f'Query {i + 1}: '
        try:
            spec = spec_from_file_location('curr', submission)
            module = module_from_spec(spec)
            spec.loader.exec_module(module)
            q = getattr(module, f'query{i + 1}')()
            ans = evaluate(q)
            def eq(i):
                _cmp = lambda a, b: any(aa == bb for aa, bb in zip(a, b))
                if type(answers[i]) is tuple:
                    return any(_cmp(ans, aa) for aa in answers[i])
                else:
                    return _cmp(ans, answers[i])
            if eq(i):
                grade += points[i]
                feedback += 'Correct.\n'
            else:
                feedback += 'Wrong Answer.\n'
                wa = True
                print('ans: ', ans, '\nsol: ', answers[i])
        except Exception as e:
            rte = True
            feedback += 'Runtime Error.\n'
            print (e)
else:
    feedback += 'No python file in submission.\n'    

max_score = sum(points)

if rte:
    feedback += ('\nPlease check for syntax errors first if you encountered runtime errors.\n' +  
        'If you believe it\'s a mistake, contact the TA at sun1226@purdue.edu for assistance.')
# output
results = {
    'output': feedback,
    'score': grade, 
    'max_score': max_score,
}

with open('/autograder/results/results.json', 'w') as res:
    json.dump(results, res)