Trying to implement OCR in bank environment but challenge is, we don't have access to internet connection for security reasons.
"Handwritten and scanned document to be digitalised"
open source like Tesseract OCR are good for normal English but most of our documents are in handwritten Arabic. I have tried Google OCR API here which has AI & ML works better with Arabic hand written and more accuracy.
I have google cloud, created storage,and uploaded hand written Arabic image in bucket then executed internal command in cloud terminal that is giving proper result.
External API Call:
curl -X POST \
-H "Authorization: Bearer "$(gcloud auth application-default print-access-token) \
-H "Content-Type: application/json; charset=utf-8" \
--data "{
'requests': [
{
'image': {
'source': {
'imageUri': 'gs://vision-api-handwriting-ocr-bucket/handwriting_image.png'
}
},
'features': [
{
'type': 'DOCUMENT_TEXT_DETECTION'
}
]
}
]
}" "https://vision.googleapis.com/v1/images:annotate"
Internal Google Cloud cmd:
gcloud ml vision detect-document "gs://vision-api-handwriting-ocr-bucket/handwriting_image.png"
The above internal google cloud cmd code works fine in my case. The same i need to implement in the local system. Is there any possibility to install same Google cloud environment and their OCR engine in local? If its priced also no problem ready to pay and buy. Resource like high configured servers, networks everything already available in our bank.
it's been a while since this was posted, but just in case, there is a solution available now.
You may check this OCR on-prem application on Google Cloud Marketplace which can be deployed as a container to any GKE cluster:
https://cloud.google.com/vision/on-prem
Related
I was trying the above api in postman. Here is the request json:
{
"input":{
"text":"Flutter is awesome!"
},
"voice":{
"languageCode":"en-gb",
"name":"en-GB-Standard-A",
"ssmlGender":"FEMALE"
},
"audioConfig":{
"audioEncoding":"MP3"
}
}
for auth, i chose Bearer in postman auth and first executed the following command in my terminal to get the token:
gcloud auth application-default print-access-token
i pasted this token in auth header, and i received the following response :
{
"error": {
"code": 403,
"message": "Your application has authenticated using end user credentials from the Google Cloud SDK or Google Cloud Shell which are not supported by the texttospeech.googleapis.com. We recommend configuring the billing/quota_project setting in gcloud or using a service account through the auth/impersonate_service_account setting. For more information about service accounts and how to use them in your application, see https://cloud.google.com/docs/authentication/. If you are getting this error with curl or similar tools, you may need to specify 'X-Goog-User-Project' HTTP header for quota and billing purposes. For more information regarding 'X-Goog-User-Project' header, please check https://cloud.google.com/apis/docs/system-parameters.",
"status": "PERMISSION_DENIED",
"details": [
{
"#type": "type.googleapis.com/google.rpc.ErrorInfo",
"reason": "SERVICE_DISABLED",
"domain": "googleapis.com",
"metadata": {
"consumer": "projects/12345678910",
"service": "texttospeech.googleapis.com"
}
}
]
}
}
I am very new to GCP in general and don't know how to navigate this issue. For additional context, i am trying to make a REST API call where i send the text and get a base64encoded string containig audio back. Any help is appreciated.
This is confusing/complex but the error is helpful:
Your application has authenticated using end user credentials from the Google Cloud SDK or Google Cloud Shell which are not supported by the texttospeech.googleapis.com.
NOTE You can try this method using Google's APIs Explorer at this link
text.synthesize.
The issue is that gcloud is an OAuth2 application and tokens issued by gcloud either using gcloud auth print-[access|identity]-token and gcloud auth application-default print-access-token are issued against a Google-managed project (that Google provides for gcloud) and -- importantly -- not one of your own projects.
Google wants to provide gcloud for its users but does not want to provide arbitrary API access (for free) to its users. Hence the "not supported" part of the error.
The solution (as described) is that you should:
Use (or create) your own Google Project
Enable the Text-to-Speech service (API) in this project
Create a Service Account and key
gcloud auth activate-service-account providing the Service Account key
gcloud auth print-access-token to get an access token to invoke the API
See the following link for the steps:
https://cloud.google.com/text-to-speech/docs/libraries
I want to build a TTS application and I am using the Google TTS engine. However, when trying the examples such as
curl -H "Authorization: Bearer "$(gcloud auth print-access-token) -H "Content-Type: application/json; charset=utf-8" --data "{
'input':{
'text':'I\'ve added the event to your calendar.'
},
'voice':{
'languageCode':'en-gb',
'name':'en-GB-Standard-A',
'ssmlGender':'FEMALE'
},
'audioConfig':{
'audioEncoding':'MP3'
}
}" "https://texttospeech.googleapis.com/v1/text:synthesize"
I get Error code 403 telling me that Cloud Text-to-Speech API has not been used in project 32555940559 before or it is disabled. Now the problem is that it mentions the project number 32555... but I want to use another project which has the TTS API switched on. Now you might suggest that I switch to the correct project, but I have done that -- when I type gcloud config list, it tells me that I am using the correct project (with project number different from the 32555...).
Where is the problem? Why is it trying to use a different project from the one that I am currently using?
You need to use "gcloud auth app-default print-access-token" (and use a dedicated service account as specified in the docs).
32555940559 is a CLOUDSDK_CLIENT_ID that comes with gcloud. And app-default has its own client_id as well. It's still now clear how Google API distinguishes between the two, unless it checks for hard-coded app-default client ID (from gcloud SDK).
It's also not clear how gcloud command line still manages to use speech API without a dedicated service account.
On a related note, since I had this same issue (API has not been used in project before or it is disabled "32555940559") in a Python virtual environment using gcloud SDKs (not curl).
Solved by running:
gcloud auth application-default login
(using gcloud auth login did not solve it).
For gcloud, the root cause might also be related to missing:
GOOGLE_CLOUD_PROJECT or GOOGLE_APPLICATION_CREDENTIALS environment variables.
Is it possible to access the Google Cloud Scheduler API using an API key?
Method: projects.locations.jobs.create
https://cloud.google.com/scheduler/docs/reference/rest/v1/projects.locations.jobs/create
I am trying to create a Job using curl:
curl -X POST \
'https://cloudscheduler.googleapis.com/v1/projects/my-project/locations/nam5/jobs?key=[MyAwesomeAPIKey]' \
-H 'Accept: application/json' \
-H 'Content-Type: application/json' \
-d '{
"name": "test-awesome-job",
"description": "My first job",
"schedule": "45 23 * * 6",
"timeZone": "utc",
"pubsubTarget": {
"topicName": "projects/my-project/topics/topic-name",
"attributes": {
"name": "39ro"
}
}
}'
but it result in a 401 Unauthorized response:
"error": {
"code": 401,
"message": "Request is missing required authentication credential. Expected OAuth 2 access token, login cookie or other valid authentication credential. See https://developers.google.com/identity/sign-in/web/devconsole-project.",
"status": "UNAUTHENTICATED"
}
The Cloud Scheduler API uses service account credentials as described in https://cloud.google.com/docs/authentication/production. As the API docs state, a limited number of GCP services support API keys and that does not include Cloud Scheduler.
If you are running the code to interact with the Cloud Scheduler API on App Engine, Cloud Functions, or Cloud Run the service account is built-in and all you need to do is grant that service account permission to interact with Cloud Scheduler via IAM.
The docs have some more streamlined information on getting set up with the Cloud Scheduler client libraries.
I get confused watching the API explorer which suggested as possible credentials
Google OAuth 2.0 or API key and from the "Help me choose" tool from Google Cloud API credentials (https://console.cloud.google.com/apis/credentials), which apparently now reports the correct solution:
For your situation you can use Application Default Credentials,
which provide a simple way to access Google APIs from App Engine or Compute Engine.
previously it showed the API keys as a possible option.
Thanks #Grayside for pointing me out!
I am using a service called Google Cloud AutoML Natural Language Processing.
Click here for more information on the product I am using:
https://beta-dot-custom-vision.appspot.com/text/overview
Everything is working fine and I need to say the product is amazing!
I managed to train a machine learning model in a couple of hours and make some predictions.
Now, I want to provide another user in my organisation with the right to access the User Interface of this product, so that this user can make his own predictions.
I connected to the Google Cloud console, created the user and assigned him three roles, which are the same that a service account would need, in order to access the service.
roles/automl.editor
roles/ml.admin
roles/storage.admin
Now, when the user tries to connect to the UI of this service, he gets a white page with the following error message:
"Exception while handling your request. When reporting refer to this issue by its tracking code tc_571165"
There is no more information, nor a trace of the issue anywhere in the google cloud console.
What am I missing? This service from Google is still in Beta, so probably it's just a bug.
Thanks for your help
In my use case, I share my vision model to my colleagues by creating a service account for them with the right role and give them this URL for prediction:
curl -X POST -H "Authorization: Bearer add_access_token " -H "Content-Type: application/json" https://automl.googleapis.com/v1beta1/projects/id_project/locations/us-central1/models/:model_idpredict -d #path_of_image_in_base64
And it works!
Hi I'm attempting to pull data from the Apple Ads API into a Google sheet, and I'm getting completely stuck on providing the security certificates. I've been able to successfully pull my data using Postman, so I'm comfortable that I can structure the request properly.
I'm trying to use URLFetchApp, but I can't see any means of including the PEM and KEY file, or even using the curl example provided by Apple of combining to the P12 file. Am I missing something here or is URLFetchApp unable to complete this?
It doesn't appear to me that this would fit into any of the existing headers for URLFetchApp https://developers.google.com/apps-script/reference/url-fetch/url-fetch-app#fetchurl-params
curl \
--cert ./<FILENAME>.p12 \
--pass <PASSWORD> \
-H "Authorization: orgId=<ORG_ID>" \
-H "Content-Type: application/json" \
-d "<CAMPAIGN_DATA_FILE>.json" \
-X POST "https://api.searchads.apple.com/api/v1/campaigns"
You're right in that Google Apps Script (GAS) does not support client-side SSL certificates in their UrlFetchApp class, which appears to be their only way to make outbound HTTP(S) requests.
Your best bet is probably to make a custom Google Apps Engine (GAE) in a language of your choice and expose an endpoint from there which when called from GAS will make a new request to your destination and provide the needed certificates. However, GAE is not free like GAS (since Google changed their cloud terms of service a couple years back), so that's something to keep in mind.