iweltag / zucker Goto Github PK

See the Sugar documentation for details here.

The API could look something like this:

async with client.bulk():
    leads_a = Lead.find(Lead.b == 3)
    leads_b = Lead.find(Lead.a == 2)
(leads_a, leads_b) = await asyncio.gather(leads_a, leads_b)

Here, we would require that any requests don't be awaited inside the context manager - that must happen afterwards, once they are all registered in some store / cache / whatever. Then, the results can be awaited and will effectively resolve alltogether.

As an alternative, this might be a bit more consice and doesn't need a context manager:

(leads_a, leads_b) = await client.bulk(
  Lead.find(Lead.b == 3),
  Lead.find(Lead.a == 2),
)

Currently, an upstream null resolves into this error:

The first idea would be to add an argument when building the field:

optional_string = model.StringField(optional=True)

Or use a wrapper:

optional_integer = model.OptionalField(model.IntegerField())

If find() gets a keyword argument fields: Iterable[str], resulting model objects should only cache those fields. There are two possibilities for what could happen when an unavailable field gets accessed:

• The corresponding value gets fetched. This could however have more implications, as it would mean that __get__ will potentially block / become awaitable depending on the pardigm.
• An exception is raised. A new method (which is sync or async, depending on context) is defined that fetches all (or only a subset) of the currently unavailable field values. It could be called something like refresh.

The latter approach is probably both easier to implement and better to use in terms of supporting both sync and async frameworks.

Sometimes happens when using the same client for a while. Traceback:

  File "/home/echobot-connector/.local/lib/python3.8/site-packages/zucker/model/view.py", line 532, in __anext__
    new_records = await self.view._module.get_client().bulk(
  File "/home/echobot-connector/.local/lib/python3.8/site-packages/zucker/client/base.py", line 619, in bulk
    await self._ensure_authentication()
  File "/home/echobot-connector/.local/lib/python3.8/site-packages/zucker/client/base.py", line 426, in _ensure_authentication
    self._finalize_authentication(auth_job_name, response_code, response_json)
  File "/home/echobot-connector/.local/lib/python3.8/site-packages/zucker/client/base.py", line 234, in _finalize_authentication
    raise ZuckerException(
zucker.exceptions.ZuckerException: authentication token renewal failed with status code 400

Relevant code:

A few places are still lacking:

• Implementing custom fields
• Differences between sync and async implementations (for example, in regards to len())

See here for a roundup of the Sugar-Specifics in upserting: https://sugarclub.sugarcrm.com/dev-club/b/dev-blog/posts/upsert-explained-by-upsert

This is a placeholder issue for the remaining field types that need to be added:

• Dropdown / Choices
• Image
• Email
• Datetime
• Date
• Boolean
• String
• Telephone number
• Multiple choice
• URL

Asynchronous views currently don't support the iteration protocol, so this isn't possible (although mentioned in the docs):

async for lead in Leads.find():
    print(lead)