Solution #8607bb8f-ff8b-4e83-a30c-14f61cfa9d6b

1def solve(input):
2    text = input.get("text", "")
3    n = len(text)
4    if n == 0:
5        return {"encoded_length": 0, "decoded": ""}
6
7    try:
8        b = text.encode("latin1")
9        counts = [0] * 256
10        for x in b:
11            counts[x] += 1
12        vals = [c for c in counts if c]
13    except:
14        counts = {}
15        for ch in text:
16            counts[ch] = counts.get(ch, 0) + 1
17        vals = list(counts.values())
18
19    m = len(vals)
20    if m == 1:
21        return {"encoded_length": n, "decoded": text}
22
23    vals.sort()
24
25    def bisect_right(a, x, lo, hi):
26        while lo < hi:
27            mid = (lo + hi) >> 1
28            if x < a[mid]:
29                hi = mid
30            else:
31                lo = mid + 1
32        return lo
33
34    total = 0
35    merged = []
36
37    while m > 1:
38        a = vals[0]
39        b = vals[1]
40        s = a + b
41        total += s
42        del vals[:2]
43        pos = bisect_right(vals, s, 0, len(vals))
44        vals[pos:pos] = [s]
45        m -= 1
46
47    return {"encoded_length": total, "decoded": text}

1def solve(input):
2    text = input.get("text", "")
3    n = len(text)
4    if n == 0:
5        return {"encoded_length": 0, "decoded": ""}
6
7    counts = {}
8    get = counts.get
9    for ch in text:
10        counts[ch] = get(ch, 0) + 1
11
12    freqs = list(counts.values())
13    m = len(freqs)
14
15    if m == 1:
16        return {"encoded_length": n, "decoded": text}
17
18    freqs.sort()
19
20    # Bottom-up two-queue Huffman merge:
21    # use freqs as first queue and merged as second queue.
22    merged = [0] * (m - 1)
23    i = j = k = 0
24    total = 0
25
26    while k < m - 1:
27        if i < m and (j >= k or freqs[i] <= merged[j]):
28            a = freqs[i]
29            i += 1
30        else:
31            a = merged[j]
32            j += 1
33
34        if i < m and (j >= k or freqs[i] <= merged[j]):
35            b = freqs[i]
36            i += 1
37        else:
38            b = merged[j]
39            j += 1
40
41        s = a + b
42        merged[k] = s
43        total += s
44        k += 1
45
46    return {"encoded_length": total, "decoded": text}