In my Flask server I am receiving a JSON-encoded parameter which is being sent via an HTTP POST from the client application.
Here is an example of what the JSON object looks like. For simplicity, I have kept only the first 2 entries, but the full object contains many more such entries.
[
{
"id": 1,
"start": 7.85,
"end": 9.813,
"text": "Θέλω να",
"words": [
"Θέλω",
"να"
],
"isBeingEditedByUser": false,
"translatedText": "I want to"
},
{
"id": 2,
"start": 9.898,
"end": 13.055,
"text": "Από κάτι το πήραν πολύ άσχημα ο οπαδός του Ολυμπιακού",
"words": [
"Από",
"κάτι",
"το",
"πήραν",
"πολύ",
"άσχημα",
"ο",
"οπαδός",
"του",
"Ολυμπιακού"
],
"isBeingEditedByUser": false,
"translatedText": "Something very bad for Olympiacos fan"
}
]
My understanding is that this JSON structure corresponds to an Array (in Javascript) or a List in Python. In this case, it is an array containing two elements, where each element is itself an object.
However, when I try to use the object on the Flask side, it seems that it has been mapped to a string (rather than a List). Is this normal behavior? I have not been able to find any documentation which states that this is the normal mapping. I would have expected the JSON object to be mapped to a Python List object instead, but this is not happening.
I know that I can use python.loads() myself to convert the string into the appropriate List structure, but I did not expect to have to do this and want to make sure that I am not misunderstanding something here.
Here is a snippet of code which shows the relevant portion in my Flask function:
#app.route('/update_SRT_file', methods=['POST'])
def update_SRT_file():
# Validate the request body contains JSON
if request.is_json:
json_obj = request.get_json()
eprint("update_SRT_file: received JSON object: ")
eprint("Type of received object is", type(json_obj));
else:
eprint("update_SRT_file: Request was not JSON ")
Here is what gets printed out:
23:12:12.771824 update_SRT_file: received JSON object:
23:12:12.771878 Type of received object is **<class 'str'>**
After more investigation, the problem was occuring because the client was JSON encoding the data twice. Upcon removing the additional encoding, it was found that now Flask correctly maps the incoming JSON Array to a python List structure.
Your POST request is sending the JSON as raw text. Flask then receives it as raw text in the request body. Some web application frameworks might automatically parse the text into a JSON-like object or data structure, but Flask does not, at least not out of the box.
Related
I'm using a LogicApp triggered by an HTTP call. The call posts a JSON message which is a single row array. I simply want to extract the single JSON object out of the array so that I can parse it but have spent several hours googling and trying various options to no avail. Here's an example of the array:
[{
"id": "866ef906-5bd8-44d8-af34-0c6906d2dfd7",
"subject": "Engagement-866ef906-5bd8-44d8-af34-0c6906d2dfd7",
"data": {
"$meta": {
"traceparent": "00-dccfde4923181d4196f870385d99cb84-52b8333f100b844c-00"
},
"timestamp": "2021-10-19T17:01:06.334Z",
"correlationId": "866ef906-5bd8-44d8-af34-0c6906d2dfd7",
"fileName": "show.xlsx"
},
"eventType": "File.Uploaded",
"eventTime": "2021-10-19T17:01:07.111Z",
"metadataVersion": "1",
"dataVersion": "1"
}]
Examples of what hasn't worked:
Parse JSON on the array, error: InvalidTemplate when specifiying an array as the schema
For each directly against the http output, error: No dependent actions succeeded.
Any suggestions would be gratefully received.
You have to paste the example that you have provided to 'Use sample payload to generate schema' in the Parse JSON Connector and then you will be able to retrieve each individual object from the sample payload.
You can extract a single JSON object from your array by using its index in square brackets. E.g., in the example below you'd need to use triggerBody()?[0] instead of triggerBody(). 0 is an index of the first element in the array, 1 - of the second, and so on.
Result:
I'm building a logic app that pulls some JSON data from a REST API, parses it with the Parse JSON block, and pushes it to Azure Log Analytics. The main problem I'm hitting is an important JSON field can either be an object or null. Per this post I changed the relevant part of my JSON schema to something like this
"entity": {"type": ["object", "null"] }
While this works, I'm now no longer to access entity later in the logic app as dynamic content. I can access all other fields parsed by the Parse JSON block downstream in the logic (that don't have nullable field). If I remove the "null" option and just have the type set to object, I can access entity in dynamic content once again. Does anyone know why this might be happening and/or how to access the entity field downstream?
Through the test, if we use "entity": {"type": ["object", "null"] }, we really cannot directly select entity in dynamic content.
But we can use the following expression to get the entity:
body('Parse_JSON')?['entity']
The test results seem to be no problem:
For a better understanding, let me cite a few more examples:
1. If your json is like this：
{
"entity": {
"testKey": "testValue"
}
}
Your expression is like this:
body('Parse_JSON')?['entity']
2. If your json is like this：
{
"test": {
"entity": {
"testKey": "testValue"
}
}
}
Your expression should like this:
body('Parse_JSON')?['test']?['entity']
I am POSTing valid JSON to a logic app and the request body is a simple set of name/value pairs in JSON format. I use parse JSON action to turn the JSON into a variable. (I am not sure if this is absolutely required, or if I can reference the http body directly without a JSON object).
[
{
"name": "fullname",
"value": "joe schmoe"
},
{
"name": "email",
"value": "joeschmoe#acme.com"
}
]
All I need to do (and this is driving me nuts) is create two variables, one containing the value of the email field and one containing the value of the fullname field.
As soon as I try to use the output value value, the logic app replaces my action with a For Each action, and then I try to to assign item().value[0] or item().value[1] to a variable without luck.
I have read dozens of examples online, but of course they all seem to be parsing JSON where there is largely unique elements in the name:value pairs.
While this is a bit of a newb question, I'm confused and need advice.
Thank you.
I used a parseJSON action to ensure I have a varaible containing the JSON array.
I then referenced the array value with (explained):
"from the body of the output from the Parse JSON action, refernce the first record in the set (fullname) and then the value of that record, the value 'joe schmoe'"
#{body('Parse_JSON')[0]['value']} (returns fullname)
Email is similar, just 2nd record in the collection:
#{body('Parse_JSON')[1]['value']} (returns email address)
I am storing c structures to couchbase, I am doing so so that I can read back these structures later and process directly, I am avoiding the steps of
1 )C structure - > JSON while storing
and
2 )JSON -> C structure while retrieving.
This is working well when I use lcb_get() and lcb_set()
But I also need have a requirement for making hits to views using the REST model and lcb_make_http_request () call.
So I was wondering how the lcb_make_http_request () will handle my non-JSON C structure , which is hex data and may have nulls in between.
Will I still be able to extract and populate my C - structure with the data that I get as HTTP response after calling lcb_make_http_request () ?
As WiredPrairie said in his comment you aren't forced to use JSON and can store C structs, but keep in mind byte order and field alignment when you are doing so.
When server detects that your data isn't in JSON format it will encode it using base64 and set meta.type to "json" when the document comes to map function.
And you will be able to emit your complete document as a value if you'd like to get the value in the HTTP stream. In case of this simple map function:
function (doc, meta) {
if (meta.type == "base64") {
emit(meta.id);
}
}
You will get response like this one (I've formatted it for clarity):
{
"total_rows": 1,
"rows": [
{
"id": "foo",
"key": "foo",
"value": "4KwuAgAAAAA="
}
]
}
It does mean that you must use some json parser to extract "value" attribute from result, decode it and you will get exactly the same bytestream, you have sent it with SET command.
I'm storing addresses data in Couchdb, and am looking for a way to get an array of just the values, instead of key: value for every record.
This is the current response:
{"total rows": 2438, "offset": 0, "rows":[
{"id": "ec5de6de2cf7bcac9a2a2a76de5738e4", "key": "user_298774", "value": {"city": "Milano", "address":"Corso Como, 42b"},
{"id": "a2a2a76de573ae4ec5de6de2cf7bcac9", "key": "user_276341", "value": {"city": "Vicenza", "address":"Via Quinto Sella, 118"}
... (etc).
]}
I only really need:
[{"city": "Milano", "address":"Corso Como, 42b"},
{"city": "Vicenza", "address":"Via Quinto Sella, 118"},
...]
I really need to minimize the usage of bandwidth that a JSON response consumes. I can't seem to find a way to transform the view into a simple array. Suggestions?
The response you are getting conforms to the Couch's REST based protocol. To reformat it two methods are provided: show functions and list functions. Basic idea is the same, but the first is suitable for retrieval documents and the list function is for you!
The list function runs the query inside the server and send the output arbitrary transformed with your JS code. API you will need is simple:
Fetch each record from the view with the getRow() function.
Export to the string (containing JSON) your JS object obj with toJSON(obj).
Send the output to the client with send(json).
If the map/reduce view URL with data is /mydb/_design/myapp/_view/mydocs-by-user and the list function name is mylist get the reformatted result to the client with the URL /mydb/_design/myapp/_list/mylist/mydocs-by-user.
Please refer to the list function documentation cited above and the chapter in the Guide for the longed tutorial.