jupyter-naas / drivers Goto Github PK

Convert all type to string to fix issue

Have a driver that will analyze and score the complexity of a text (Flesh)

Input:

• API key
• Pipeline name
• Column names

Model:
get status of each column in pipeline weighted by % of chance of success (inside the column property)

Ouput:
Dataframe/csv/email

https://naas.canny.io/private/p/sales-reprendreremettre-en-marche-lemail-journalier-qui-donne-letat-du-pipeline

→ Solution to be studied before implementation.

Newsapi driver Type error

Error: TypeError: get() got multiple values for argument 'fields'

going through the code, get method has "image" specified twice

def get(self, fields=["image", "title", "image", "link"], limit=20, **kargs):

Following work from @gagan3012 we would like to make sure HuggingFace is properly identified as integration in Naas and not as a feature provided by Naas drivers.

What is expected?

A page or a database entry fetched by notion driver populates and persists "Relation" type page property.

What is observed?

"Relation" type page property is not populated in naas notion driver. So when I update any of my databases with any other field (eg. select, rich_text, url etc) all "Relation" type properties get set to "" (empty value).

🧑‍💻 👉 Each time an entry in my database is updated usin naas notion driver, the relation property get deleted. 😭 I ran the script in a cron job. So in a few days all the reation information in my database dissapeared. 💔

Additional Information

1. when a database is fetched, the relation properties in page looks like this

COLUMN_NAME: !!python/object:naas_drivers.tools.notion.PagePropertyRelation
  id: Sqt%7D
  relation: []
  type: relation

ie the relation field is []
2. what i could find in notion api documentation
- https://developers.notion.com/reference/property-object#database-properties
- https://developers.notion.com/reference/property-object#relation-configuration
3. looks like something is broken here:

Got a warning at import of naas_drivers related to Slack.

Due to a bug calling, naas_drivers would also open a port for streamlit that can cause safety issues

This will be fixed in the next PR

As of today, the naas_drivers folder has different subfolders

• Input
• InputOuput
• Output

We want to put everything under one "tools" folder that will make the development and search of integrations easier.

The goal is to get some insight about the usage of our LinkedIn API. This will allow us to understand how the api is used, detect potential errors, make sure the usage stays "correct" (in term of traffic) during the beta testing phase.

Like GSheet airtable should allow replacing all the data

List of formulas we could develop to update current drivers

Connect

token = "*********"
naas_drivers.notion.connect(token)

Get page

url = "*********"
naas_drivers.notion.connect(token).page.get(url)

Get database

url = "*********"
naas_drivers.notion.connect(token).database.get(url)

Create page

parent_id = "*********" #can be id of wokspace or id or parent page
naas_drivers.notion.connect(token).page.create(parent_id, properties, content)

Add properties

properties.name = "The name of my page"
properties.created_date = "14/09/2021"

content.heading1 = "My heading 1"
content.heading2 = "My heading 2"

To be developped further @axelearning @FlorentLvr

The first point of prediction should be the last one from the current dataset

I have encoutered the following error message when running the template "LinkedIn - Get posts stats from profile"

remove install version

Link to the template: jupyter-naas/awesome-notebooks#201

Run query in files stored in AWS bucket S3
Today, I store my petabytes of data in a bucket S3, I would be able to query those files directly through naas.

TBD

I noticed when using linkedin.connect(LI_AT, JSESSIONID).profile.get_posts_stats(PROFILE_URL) I was having data quality issues where I had 0 views, which is unlikely on this platform.

When clicking into the provided URLs I was taken to the following screen:

Upon further inspection, it seems that 1) this function struggles to pick up LinkedIn poll posts, and 2) when sharing posts from other users, it attributes the shared post as your own with 0 for all stats. Examples of such posts are below:

A model and a template exist, need to be packaged into a driver.
Ask for more details: [email protected]

With naas_drivers.vpn.connect() I can login to a company VNP to send and get remote files.

Install pymongo[srv] to connect to mongoDB in driver

Add the current company in the "get_profile" driver.

Today, it's possible from Google colab to create an integration with Streamlit (dashboarding/apps) using a "ngrok".
This could be also integrated in Naas as a driver.
import naas
naas.streamlit.add(source="mygraph.html", target="streamlit account")

https://naas.canny.io/requests/p/create-streamlit-driver

Following discussion with D.Rothman, Gagan Bhatia

Create new function to get polls result in LinkedIn driver:

linkedin.connect().post.get_polls()

The name is Zappier instead of Zapier

Add the current profile picture to the get profile function.

→ potentially scrapping with beautiful soup

Delete this line to get rid of error message.
Seen with @Dr0p42

if not given full link for id:

no schema supplied error

if full link is given:

JSONDecodeError: Expecting value: line 1 column 1 (char 0)
in data = data place

User: alexandre inukai

"I can't get my data using the notion driver.
Error: ""message", "There was an error (400) submitting the request."
HTTPError: Invalid input."
Token is correct as it is working with another unofficial api (notion automation).
I try to get a notion table which is in table view.
Same error when trying to get the content of a page.
No idea where it can come from as it is working for my personal notion account.
Thanks
"
March 25, 2021

https://github.com/jupyter-naas/drivers/tree/master/naas_drivers

As a user, I want to be able to see the requirements from the README to be able to install it locally.

OS: Windows...
Package: Python version /

Function : github.connect(GITHUB_TOKEN).repos.get_stargazers(REPO_URL)

Add a driver to support the taggun receipt ocr API

https://www.taggun.io/

Should allow requesting both simple and verbose results

eventually should also support resubmitting for training purposes

Environment

• Windows 8.1
• Conda-4.10.3
• python 3.8

Error : Command errored out with exit status 1: python setup.py egg_info check the logs for more info

Error elaborated: UnicodeDecodeError: 'charmap' codec can't decode byte 0x8f in position 1130: character maps to

Setup.py line 4 long_description = fh.read() is causing this error

wait randomly between 2 and 4 seconds between each requests

please see below error message