我需要帮助从Python的嵌入式字典中提取最大值

Question

So I have seen a couple of posts on this but they don't line up with my needs because my dict is names, the other posts the keys all have the same string label (key) and I'm not sure how to work around that. I have a defaultdict of dicts with the values being their similarity vectors. so effectively what its doing is the first key is the person to whom all the other keys are comparing and the value is how similar they are. I want to find the most similar.

finalDict = defaultdict(dict)

for s in meanDict.keys():
    spk1 = s
    vecSpk1 = meanDict[spk1]
    for t in meanDict.keys():
        if t != spk1:
            spk2 = t
            vecSpk2 = meanDict[spk2]
            sim = Similarity(vecSpk1,vecSpk2)
            finalDict[spk1][spk2] = sim

This is what I have tried so far but i get the error code: TypeError: unhashable type: 'dict'

for s in finalDict.keys():
    speaker = finalDict[s]
    max([float(d[speaker]) for d in finalDict.values()]) 

one sample entry looks like this:

defaultdict(<type 'dict'>, {
    'Rand Paul': {
        'Rick Santorum': 0.9915748250754691, 
        'Joseph Biden': 0.9941726131674848, 
        'Lindsey Graham': 0.9976981759787777, 
        'Mike Pence': 0.9898898183716316, 
        'Hillary Clinton': 0.9909482785262219, 
        'John Kasich': 0.9891372990166918, 
        'Barack Obama': 0.9914114570907536, 
        'Marco Rubio': 0.9806612828270346, 
        'Jim Webb': 0.9863228519191483, 
        'Bernie Sanders': 0.9940352419978636, 
        'Ted Cruz': 0.993712852439669, 
        'Lincoln Chafee': 0.9763407586067823
    }
}

Answer 1

To fix your program, you can change the last line to:

max([d for d in speaker.values()])

However, there are ways to further improve your code. For instance, consider using the items() method of dicts and the key attribute of the max function:

for s, speakers in finalDict.items():
    max(speakers.items(), key=lambda x: x[1])