在python中计算json值平均值

Question

how can I calculate in python the values JSON file in the following example:

"items": [
         {
             "start": "0.6",
             "end": "0.9",
             "alter": [
                 {
                     "conf": "0.6",
                     "content": ""
                 }
             ],
             "type": "pron"
         },
     ]

Answer 1

import json

with open("./file.json") as f:
    dict_data = json.load(f)  # passing file object and will return json in dictionary datatype
confidences = [float(i['alternatives'][0]['confidence']) for i in dict_data['items']]
confidence_avg = sum(confidences) / len(confidences)
print(confidence_avg)

Output:

0.8534666666666667

Answer 2

For starters, your JSON file is missing the first and last curly brackets, so I've added them manually. Without them, it is not valid JSON.

Use json.loads to parse the JSON string and return a dict .

The confidence values are stored as strings, so they need to be transformed to float s.

Add them one by one and divide by the number of confidence values. In this case we assume each item has only 1.

import json

json_str = r"""{
    "items": [
        {
            "start_time": "0.0",
            "end_time": "0.46",
            "alternatives": [
                {
                    "confidence": "0.9534",
                    "content": "رسالة"
                }
            ],
            "type": "pronunciation"
        },
        {
            "start_time": "0.46",
            "end_time": "0.69",
            "alternatives": [
                {
                    "confidence": "0.6475",
                    "content": "اللغة"
                }
            ],
            "type": "pronunciation"
        },
        {
            "start_time": "0.69",
            "end_time": "1.23",
            "alternatives": [
                {
                    "confidence": "0.9595",
                    "content": "العربية"
                }
            ],
            "type": "pronunciation"
        }
    ]
}"""

items = json.loads(json_str)["items"]

average = 0
for item in items:
    confidence = float(item["alternatives"][0]["confidence"])
    average += confidence

average /= len(items)

print(average)

Output:

0.8534666666666667