JSON¶
概述¶
JSON is built on two structures:
- collections of key/value pairs
- lists of values
Python includes a json module which helps you encode and decode JSON.
Create JSON Demo¶
create_json_from_dict.py¶
import json
# Create a dictionary object
person_dict = {'first': 'Christopher', 'last':'Harrison'}
# Add additional key pairs to dictionary as needed
person_dict['City']='Seattle'
# Convert dictionary to JSON object
person_json = json.dumps(person_dict)
# Print JSON object
print(person_json)
create_json_with_list.py¶
import json
# Create a dictionary object
person_dict = {'first': 'Christopher', 'last':'Harrison'}
# Add additional key pairs to dictionary as needed
person_dict['City']='Seattle'
# Create a list object of programming languages
languages_list = ['CSharp','Python','JavaScript']
# Add list object to dictionary for the languages key
person_dict['languages']= languages_list
# Convert dictionary to JSON object
person_json = json.dumps(person_dict)
# Print JSON object
print(person_json)
create_json_with_nested_dict.py¶
import json
# Create a dictionary object
person_dict = {'first': 'Christopher', 'last':'Harrison'}
# Add additional key pairs to dictionary as needed
person_dict['City']='Seattle'
# create a staff dictionary
# assign a person to a staff position of program manager
staff_dict ={}
staff_dict['Program Manager']=person_dict
# Convert dictionary to JSON object
staff_json = json.dumps(staff_dict)
# Print JSON object
print(staff_json)
Read JSON Demo¶
read_json.py¶
# This code will show you how to call the Computer Vision API from Python
# You can find documentation on the Computer Vision Analyze Image method here
# https://westus.dev.cognitive.microsoft.com/docs/services/5adf991815e1060e6355ad44/operations/56f91f2e778daf14a499e1fa
# Use the requests library to simplify making a REST API call from Python
import requests
# We will need the json library to read the data passed back
# by the web service
import json
# We need the address of our Computer vision service
vision_service_address = "https://canadacentral.api.cognitive.microsoft.com/vision/v2.0/"
# Add the name of the function we want to call to the address
address = vision_service_address + "analyze"
# According to the documentation for the analyze image function
# There are three optional parameters: language, details & visualFeatures
parameters = {'visualFeatures':'Description,Color',
'language':'en'}
# We need the key to access our Computer Vision Service
subscription_key = "cf229a23c3054905b5a8ad512edfa9dd"
# Open the image file to get a file object containing the image to analyze
image_path = "./TestImages/PolarBear.jpg"
image_data = open(image_path, 'rb').read()
# According to the documentation for the analyze image function
# we need to specify the subscription key and the content type
# in the HTTP header. Content-Type is application/octet-stream when you pass in a image directly
headers = {'Content-Type': 'application/octet-stream',
'Ocp-Apim-Subscription-Key': subscription_key}
# According to the documentation for the analyze image function
# we use HTTP POST to call this function
response = requests.post(address, headers=headers, params=parameters, data=image_data)
# Raise an exception if the call returns an error code
response.raise_for_status()
# Display the JSON results returned
results = response.json()
print(json.dumps(results))
print('requestId')
print (results['requestId'])
print('dominantColorBackground')
print(results['color']['dominantColorBackground'])
print('first_tag')
print(results['description']['tags'][0])
for item in results['description']['tags']:
print(item)
print('caption text')
print(results['description']['captions'][0]['text'])
read_key_pair.py¶
# This code will show you how to call the Computer Vision API from Python
# You can find documentation on the Computer Vision Analyze Image method here
# https://westus.dev.cognitive.microsoft.com/docs/services/5adf991815e1060e6355ad44/operations/56f91f2e778daf14a499e1fa
# Use the requests library to simplify making a REST API call from Python
import requests
# We will need the json library to read the data passed back
# by the web service
import json
# We need the address of our Computer vision service
vision_service_address = "https://canadacentral.api.cognitive.microsoft.com/vision/v2.0/"
# Add the name of the function we want to call to the address
address = vision_service_address + "analyze"
# According to the documentation for the analyze image function
# There are three optional parameters: language, details & visualFeatures
parameters = {'visualFeatures':'Description,Color',
'language':'en'}
# We need the key to access our Computer Vision Service
subscription_key = "xxxxxxxxxxxxxxxxxxxx"
# Open the image file to get a file object containing the image to analyze
image_path = "./TestImages/PolarBear.jpg"
image_data = open(image_path, 'rb').read()
# According to the documentation for the analyze image function
# we need to specify the subscription key and the content type
# in the HTTP header. Content-Type is application/octet-stream when you pass in a image directly
headers = {'Content-Type': 'application/octet-stream',
'Ocp-Apim-Subscription-Key': subscription_key}
# According to the documentation for the analyze image function
# we use HTTP POST to call this function
response = requests.post(address, headers=headers, params=parameters, data=image_data)
# Raise an exception if the call returns an error code
response.raise_for_status()
# Display the raw JSON results returned
results = response.json()
print(json.dumps(results))
# print the value for requestId from the JSON output
print()
print('requestId')
print (results['requestId'])
read_key_pair_list.py¶
# This code will show you how to call the Computer Vision API from Python
# You can find documentation on the Computer Vision Analyze Image method here
# https://westus.dev.cognitive.microsoft.com/docs/services/5adf991815e1060e6355ad44/operations/56f91f2e778daf14a499e1fa
# Use the requests library to simplify making a REST API call from Python
import requests
# We will need the json library to read the data passed back
# by the web service
import json
# We need the address of our Computer vision service
vision_service_address = "https://canadacentral.api.cognitive.microsoft.com/vision/v2.0/"
# Add the name of the function we want to call to the address
address = vision_service_address + "analyze"
# According to the documentation for the analyze image function
# There are three optional parameters: language, details & visualFeatures
parameters = {'visualFeatures':'Description,Color',
'language':'en'}
# We need the key to access our Computer Vision Service
subscription_key = "xxxxxxxxxxxxxxxxxxxxxxx"
# Open the image file to get a file object containing the image to analyze
image_path = "./TestImages/PolarBear.jpg"
image_data = open(image_path, 'rb').read()
# According to the documentation for the analyze image function
# we need to specify the subscription key and the content type
# in the HTTP header. Content-Type is application/octet-stream when you pass in a image directly
headers = {'Content-Type': 'application/octet-stream',
'Ocp-Apim-Subscription-Key': subscription_key}
# According to the documentation for the analyze image function
# we use HTTP POST to call this function
response = requests.post(address, headers=headers, params=parameters, data=image_data)
# Raise an exception if the call returns an error code
response.raise_for_status()
# Display the JSON results returned in their raw JSON format
results = response.json()
print(json.dumps(results))
# Print out all the tags in the description
print()
print('all tags')
for item in results['description']['tags']:
print(item)
# print out the first tag in the description
print()
print('first_tag')
print(results['description']['tags'][0])
read_subkey.py
# This code will show you how to call the Computer Vision API from Python
# You can find documentation on the Computer Vision Analyze Image method here
# https://westus.dev.cognitive.microsoft.com/docs/services/5adf991815e1060e6355ad44/operations/56f91f2e778daf14a499e1fa
# Use the requests library to simplify making a REST API call from Python
import requests
# We will need the json library to read the data passed back
# by the web service
import json
# We need the address of our Computer vision service
vision_service_address = "https://canadacentral.api.cognitive.microsoft.com/vision/v2.0/"
# Add the name of the function we want to call to the address
address = vision_service_address + "analyze"
# According to the documentation for the analyze image function
# There are three optional parameters: language, details & visualFeatures
parameters = {'visualFeatures':'Description,Color',
'language':'en'}
# We need the key to access our Computer Vision Service
subscription_key = "xxxxxxxxxxxxxxxxxxxxxxx"
# Open the image file to get a file object containing the image to analyze
image_path = "./TestImages/PolarBear.jpg"
image_data = open(image_path, 'rb').read()
# According to the documentation for the analyze image function
# we need to specify the subscription key and the content type
# in the HTTP header. Content-Type is application/octet-stream when you pass in a image directly
headers = {'Content-Type': 'application/octet-stream',
'Ocp-Apim-Subscription-Key': subscription_key}
# According to the documentation for the analyze image function
# we use HTTP POST to call this function
response = requests.post(address, headers=headers, params=parameters, data=image_data)
# Raise an exception if the call returns an error code
response.raise_for_status()
# Display the raw JSON results returned
results = response.json()
print(json.dumps(results))
# Print the value for dominantColorBackground from the color keys
print()
print('dominantColorBackground')
print(results['color']['dominantColorBackground'])