I've been asked to swap over from urllib2 to the requests library because the library is simpler to use and doesn't cause exceptions.
I can get the HTTP error code with response.status_code but I don't see a way to get the error message for that code. Normally, I wouldn't care, but I'm testing an API and the string is just as important.
Does anybody know of a simple way to get that string? I'm expecting 2 pieces something like:
'400':'Bad Request'
This is NOT a DUPLICATE
Some of the codes being returned have unique strings being sent by the application that I am testing. These strings cannot be looked up using this method: requests.status_codes._codes[error][0] Since the string is dynamically coming from the back end server. I was able to get this information using urllib using this method:
import urllib2
...
opener = urllib2.build_opener(urllib2.HTTPSHandler(context=ctx))
except urllib2.HTTPError as err:
try: error_message = err.read()
...
The question now is... is there a method of getting the dynamic http error string? Thanks so much for being patient. The previous issue was closed so quickly I never got a chance to look at the answer, test it and re-ask by cleaning up the description.
response = requests.get(url)
error_message = response.reason
In HTTPResponse there's a reason attribute that returns the reason phrase from the response's status line. In the requests library the requests.Response class has an equivalent reason attribute that returns the same thing. Both should return the information from the response, not a fixed string based on the code.
Related
I am a new programmer and I'm learning the request module. I'm stuck on the fact that I don't know how to get a specific part of a json response, I think it's called a header? or its the thing inside of a header? I'm not sure. But the API returns simple json code. This is the api
https://mcapi.us/server/status?ip=mc.hypixel.net
for more of a example, lets say it returns this json code from the api
{"status":"success","online":true"}
And I wanted to get the "online" response, how would I do that?
And this is the code im currently working with.
import requests
def main():
ask = input("IP : ")
response = requests.get('https://mcapi.us/server/status?ip=' + ask)
print(response.content)
main()
And to be honest, I don't even know if this is json. I think it is but the api page says its cors? if it isn't I'm sorry.
In your example you have a dictionary with key "online"
You need to parse it first with .json() and then you can get it in form dict[key]
In your case
response = requests.get('https://mcapi.us/server/status?ip=' + ask).json()
print(response["online"])
or in case of actual content
response = requests.get('https://mcapi.us/server/status?ip=' + ask).json()
print(response["content"])
I've just discovered something strange. When downloading data from facebook with GET using the requests 2.18.4 library, I get error when I just use
requests.get('https://.../{}/likes?acces_token={}'.format(userID,token))
into which I parse the user ID and access - the API does not read the access token correctly.
But, it works fine as
requests.get('https://../{}'.format(userID), params={"access_token":token})
Or it works when I copy paste the values in the appropriate fields by hand in the python console.
So my hypothesis is that it has something to with how the token string got parsed using the params vs the string. But what I don't understand at all, why would that be the case? Or is ? character somehow strange in this case?
Double check if both the URLs are the same (in your post they differ by the /likes substring).
Then you can check how the library requests concatenated parameters from the params argument:
url = 'https://facebook.com/.../{}'.format(userID)
r = requests.Request('GET', url, params={"access_token":token})
pr = r.prepare()
print pr.url
I am trying to extract the URL of specific articles from NYT API.
This is my code:
import requests
for i in range(0,100):
page=str(i)
r = requests.get("http://api.nytimes.com/svc/search/v2/articlesearch.json?begin_date=20100101&q=terrorist+attack&page="+page+"&api-key=***")
data = r.json()
article = data['response']['docs']
for url in article:
print(url["web_url"])
After printing the first 20 URL it gives me this error
KeyError: 'response'
however by checking random pages the key 'response' is present in any of them. What can I do to print all the URLs from the next 88 pages?
I ran into a similar problem. You might be requesting faster than the allowed limit of 5 per second. In that case, the NYT server is going to hit you with an error message, so there will be no 'response' key. I would suggest printing out the keys from every GET request using something like:
print(dict.keys(data))
If you keep seeing 'message' as one of your keys, then you know that you're probably requesting too fast. So just put in a time.sleep(0.5) to slow things down and you should be good.
You are assuming that there are at least 101 pages to make requests (0 to 100).
If you make a request to page 100, do you still get the same JSON structure with a response key?
What you should instead use is a while loop that breaks when you get a KeyError.
This is my first couple of weeks coding; apologies for a basic question.
I've managed to parse the 'WorldNews' subreddit json, identify the individual children (24 of them as I write) and grab the titles of each news item. I'm now trying to create an array from these news titles. The code below does print the fifth title ([4]) to command line every 2-3 attempts (otherwise provides the error below). It will also not print more than one title at a time (for example if I try[2,3,4] I will continuously get the same error).
The error I get when doesn't compile:
in <module> Children = theJSON["data"]["children"] KeyError: 'data'
My script:
import requests
import json
r = requests.get('https://www.reddit.com/r/worldnews/.json')
theJSON = json.loads(r.text)
Children = theJSON["data"]["children"]
News_list = []
for post in Children:
News_list.append (post["data"]["title"])
print News_list [4]
I've managed to find a solution with the help of Eric. The issue here was in fact not related to the key, parsing or presentation of the dict or array. When requesting a Url from reddit and attempting to print the json string output we encounter an HTTP Error 429. Fixing this is simple. The answer was found on this redditdev thread.
Solution: by adding an identifier for the device requesting the Url ('User-agent' in header) it runs smoothly and works every time.
import requests
import json
r = requests.get('https://www.reddit.com/r/worldnews.json', headers = {'User-agent': 'Chrome'})
theJSON = json.loads(r.text)
print theJSON
This means that the payload you got didn't have a data key in it, for whatever reason. I don't know about Reddit's JSON API; I tested the request and saw that you were using the correct keys. The fact that you say your code works every few times tells me that you're getting a different response between requests. I can't reproduce it, I tried making the request over and over and checking for the correct response. If I had to guess why you'd get something different I'd say it'd have to be either rate limiting or a temporary 503 (Reddit having issues.)
You can guard against this by either catching the KeyError or using the .get method of dictionaries.
Catching KeyError:
try:
Children = theJSON["data"]["children"]
except KeyError:
print 'bad payload'
return
Using .get:
Children = theJSON.get("data", {}).get("children")
if not Children:
print 'bad payload'
return
I'm trying to understand what exactly I'm getting back when I make a POST request using the Requests module — is it always JSON? Seems like every response I get appears to be JSON, but I'm not sure.
Where r is my response object, when I do:
print r.apparent_encoding
It always seems to return ascii
And when I try type():
>>>print type(r)
<class 'requests.models.Response'
I pasted the output from print r.text into a JSON validator, and it reported no errors. So should I assume Requests is providing my with JSON objects here?
A response can be anything. If you've posted to a REST endpoint, it will usually respond with JSON. If so, requests will detect that and allow you to decode it via the .json() method.
But it's perfectly possible for you to post to a normal web URL, in effect pretending to be a browser, and unless the server is doing something really clever it will just respond with the standard HTML it would serve to the browser. In that case, doing response.json() will raise a ValueError.
No, the response text for a POST request is totally up to the web service. A good REST API will always respond with JSON, but you will not always get that.
Example
A common pattern in PHP is
<?php
$successful_whatever = false;
if (isset($_POST['whatever'])) {
# put $_POST['whatever'] in a database
$successful_whatever = true;
}
echo $twig->render('gallery.twig',
array('successful_whatever' => $successful_whatever));
?>
As you can see the response text will be a rendered template (HTML). I'm not saying it is good, just that it is common.