Speechtext #15

JJneid · 2022-12-28T01:06:13Z

issues addressed:
issue #14
issue #7
issue #6
issue #11
issue #13

eff-kay · 2022-12-28T01:14:05Z

README.md

+# SLASHML Python client

-to be updated to guide users
+This is a Python client for SLASHML. It lets you run transcription jobs from your Python code or Jupyter notebook. Do a transcription job with three lines of code


add url for slashml

The readme should be generic, so it doesn't run transcrption jobs. It runs machine learning jobs.

Then you can give an example of how to do transcription as an example.

eff-kay · 2022-12-28T01:15:49Z

src/LICENSE.txt

@@ -0,0 +1,21 @@
+MIT License


the only license we need should be at the root, so remove all the licenses inside the folders.

you have to push the chnages, the license at src/LICENSE.txt still exists in this PR, we wan't to remove it.

eff-kay · 2022-12-28T01:16:29Z

src/slashml.py

+    HEADERS:dict = {}
+    ## add the api key to sys path envs 
+    def __init__(self,API_KEY: str = None):
+        self.API_KEY=None


self.API_KEY is not used.

change is not pushed

eff-kay · 2022-12-28T01:17:16Z

src/slashml.py

+           # print("Auth with environment variable SLASHML_API_KEY")
+        else:
+            self.HEADERS=None 
+            print("No Auth, there are certain limites to the usage")


No API key provided, there are limits to the usage.

eff-kay · 2022-12-28T01:18:18Z

src/slashml.py

+        ]
+        response = requests.post(self.SLASHML_UPLOAD_URL,
+                                headers=headers,files=files)
+        return  response.json()["upload_url"]


there is a possibility that you receive 400 or 500 when uploading the file. When that happens the response.json() might not contain upload_url.

i added conditions on this, if the response is success get the upload URL.

eff-kay · 2022-12-28T01:19:45Z

src/slashml.py

+        response = requests.post(self.SLASHML_TRANSCRIPT_URL, headers=headers, data=payload)
+        # Check the status code of the response
+        if response.status_code == 200:
+            return response.json()["id"]


this is correct, but are you sure you just want to return the ID from transcribe?

for now yes,

eff-kay · 2022-12-28T01:20:27Z

src/slashml.py

+        if response.status_code == 200:
+            return response.json()["id"]
+
+        elif response.status_code == 429:


Don't check for all invidual error codes. Just pass the error to the user as a reponse. Also you want to pass the error msg.

eff-kay · 2022-12-28T01:20:59Z

src/slashml.py

+                                headers=headers,files=files)
+        return  response.json()["upload_url"]
+
+    def transcribe(self,upload_url:str,  service_provider: str,header=None ):


whats the output of transcription?

eff-kay · 2022-12-28T01:21:09Z

src/slashml.py

+        else:
+            return "UNKNOWN ERROR"
+
+    def status(self,job_id:str, service_provider: str ,header=None):


whats the output of status?

eff-kay · 2022-12-28T01:21:26Z

src/test_speechtotext_1.py

@@ -0,0 +1,21 @@
+import slashml


remove this file, or convert it into tests

eff-kay · 2022-12-28T01:21:36Z

src/test_speechtotext_1.py

+# Initialize SlashML 
+speect_to_text = slashml.SpeechToText()
+# optional local file to upload, if not an accessible url
+file_location="/Users/JJneid/Desktop/SlashMl/Benchmarking/podcast1/podcast1_long_trim.mp3"


this location is specific to your computer.

eff-kay · 2022-12-29T08:49:38Z

README.md

+[SlashML](https://www.slashml.com/)
+# Introduction
+## Overview 
+This is a Python client (SDK) for SlashML. It allows the user to use the apps available and active in [SlashML-Dashboard](https://www.slashml.com/dashboard). 


we call them endpoints or services, not apps.

eff-kay · 2022-12-29T08:50:13Z

README.md

+# Introduction
+## Overview 
+This is a Python client (SDK) for SlashML. It allows the user to use the apps available and active in [SlashML-Dashboard](https://www.slashml.com/dashboard). 
+The apps can be together in the same code. For example, if the user wants to transcribe an audio file and get a summary, they can call speechtotext followed by summarization.


Suggested change

The apps can be together in the same code. For example, if the user wants to transcribe an audio file and get a summary, they can call speechtotext followed by summarization.

The endpoints can be combined together in any order. For example, if the user wants to transcribe an audio file and get a summary, they can call `speechtotext` followed by summarization.

eff-kay · 2022-12-29T08:51:42Z

README.md

+The apps can be together in the same code. For example, if the user wants to transcribe an audio file and get a summary, they can call speechtotext followed by summarization.

-to be updated to guide users
+State-of-the-art AI models from several service providers are available. At SlashML, we also do the benchmarking on these models for the user. This will give them an idea of the best service provider for their application. For the full list of the available models thru SlashML, click [here](##availlable-service-providers)


eff-kay · 2022-12-29T08:52:27Z

README.md

+The apps can be together in the same code. For example, if the user wants to transcribe an audio file and get a summary, they can call speechtotext followed by summarization.

-to be updated to guide users
+State-of-the-art AI models from several service providers are available. At SlashML, we also do the benchmarking on these models for the user. This will give them an idea of the best service provider for their application. For the full list of the available models thru SlashML, click [here](##availlable-service-providers)


this will give who an idea for the best service provider. The user is reading the email. Write in third person, not second person.

eff-kay · 2022-12-29T08:52:49Z

README.md

+
+
+
+to lets you run transcription jobs from your Python code or Jupyter notebook. The transcription can be done with three lines of code


Capitalize to

eff-kay · 2022-12-29T08:53:24Z

README.md

+
+
+
+to lets you run transcription jobs from your Python code or Jupyter notebook. The transcription can be done with three lines of code


The first line of this paragraph does not make sense.

eff-kay · 2022-12-29T08:53:59Z

README.md

+
+speect_to_text = slashml.SpeechToText()
+transcribe_id= speect_to_text.transcribe(audio_url,service_provider="aws")
+status=speect_to_text.status(transcribe_id,service_provider=service_provider)


This code will not work, you need to wait for the status to turn to COMPLETED.

eff-kay · 2022-12-29T08:55:38Z

README.md

+```
+
+## Set up and usage
+There is a daily limit (throttling) on the number of calls the user performs,  the code runs without specifying a token (API key),  the throttling kicks in and prevents new jobs after exceeding 10 calls per minute. 


Suggested change

There is a daily limit (throttling) on the number of calls the user performs, the code runs without specifying a token (API key), the throttling kicks in and prevents new jobs after exceeding 10 calls per minute.

There is a daily limit (throttling) on the number of calls the user performs. The code can run without specifying a token (API key). However, the throttling will limit job requests that exceed 10 calls per minute.

eff-kay · 2022-12-29T08:55:57Z

README.md

+## Set up and usage
+There is a daily limit (throttling) on the number of calls the user performs,  the code runs without specifying a token (API key),  the throttling kicks in and prevents new jobs after exceeding 10 calls per minute. 
+
+If the user intends on using the service more frequently, it is recommended to generate an token or API key from [SlashML-Dashboard](https://www.slashml.com/dashboard). This way, the throttling limit will increase.


generate a token or an API key

eff-kay · 2022-12-29T08:56:53Z

README.md

+
+If the user intends on using the service more frequently, it is recommended to generate an token or API key from [SlashML-Dashboard](https://www.slashml.com/dashboard). This way, the throttling limit will increase.
+
+Sign up and Grab your token from [SlashML-Dashboard](https://www.slashml.com/dashboard)>settings> new api key and authenticate by setting it as an environment variable (or when you initialize the service, see Quickstart tutorial):


This can be a new subsection. Also, reword this

eff-kay · 2022-12-29T08:58:14Z

README.md

+
+In your terminal
+```
+export SLASHML_API_KEY=[token]


this only works for macs and unix, will not work in windows. Specify how to set it in the terminal in different OSes.

eff-kay · 2022-12-29T08:58:59Z

README.md

+
+```
+
+If the user preferes using the API calls directly, the documentation is available [here](https://www.slashml.com/dashboard).


This one seems a sudden jump, in thought Should be a separate section.

eff-kay · 2022-12-29T08:59:18Z

README.md

+AWS
+Whisper (OpenAI)
+
+### Summarization


Remove summarization from this PR.

eff-kay · 2022-12-29T08:59:51Z

README.md

+hugging-face summarization based on Meta... include links
+# Quickstart tutorial 
+
+## Introduction


The documentation is not finalized, remove the extra contents.

JJneid added 2 commits December 27, 2022 19:48

speech to text sdk

fcfcb6c

speech to text with license

58036c1

JJneid requested a review from eff-kay December 28, 2022 01:06

eff-kay reviewed Dec 28, 2022

View reviewed changes

changing name of import

af50882

eff-kay reviewed Dec 28, 2022

View reviewed changes

src/test_speechtotext_1.py

@@ -0,0 +1,21 @@

import slashml

Copy link

Contributor

eff-kay Dec 28, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

remove this file, or convert it into tests

eff-kay reviewed Dec 28, 2022

View reviewed changes

readme continous update

a25a1cc

eff-kay reviewed Dec 29, 2022

View reviewed changes

README.md

AWS

Whisper (OpenAI)

### Summarization

Copy link

Contributor

eff-kay Dec 29, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Remove summarization from this PR.

eff-kay reviewed Dec 29, 2022

View reviewed changes

eff-kay merged commit a25a1cc into main Jan 14, 2023

	The apps can be together in the same code. For example, if the user wants to transcribe an audio file and get a summary, they can call speechtotext followed by summarization.
	The endpoints can be combined together in any order. For example, if the user wants to transcribe an audio file and get a summary, they can call `speechtotext` followed by summarization.




		to lets you run transcription jobs from your Python code or Jupyter notebook. The transcription can be done with three lines of code

	There is a daily limit (throttling) on the number of calls the user performs, the code runs without specifying a token (API key), the throttling kicks in and prevents new jobs after exceeding 10 calls per minute.
	There is a daily limit (throttling) on the number of calls the user performs. The code can run without specifying a token (API key). However, the throttling will limit job requests that exceed 10 calls per minute.


		If the user intends on using the service more frequently, it is recommended to generate an token or API key from [SlashML-Dashboard](https://www.slashml.com/dashboard). This way, the throttling limit will increase.

		Sign up and Grab your token from [SlashML-Dashboard](https://www.slashml.com/dashboard)>settings> new api key and authenticate by setting it as an environment variable (or when you initialize the service, see Quickstart tutorial):


		```

		If the user preferes using the API calls directly, the documentation is available [here](https://www.slashml.com/dashboard).

Speechtext #15

Speechtext #15

Uh oh!

Conversation

JJneid commented Dec 28, 2022

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

eff-kay Dec 29, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

eff-kay Dec 29, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

eff-kay Dec 29, 2022 •

edited

Loading

eff-kay Dec 29, 2022 •

edited

Loading