Draft: Speed improvements #659
Conversation
|
@arthanson the size of this PR may have gotten a little bit out of hand. As a summary here is what's been implemented so far:
Please let me know what you think and what you'd like to possibly see as individual PR. |
| @@ -109,3 +109,5 @@ | |||
| "wireless.WirelessLANGroup": None, | |||
| "wireless.wirelesslink": None, | |||
| } | |||
|
|
|||
| TYPE_CONTENT_MAPPER = {v: k for k, v in CONTENT_TYPE_MAPPER.items() if v is not None} | |||
There was a problem hiding this comment.
What is this for? I don't see it used anywhere?
| @@ -151,6 +151,7 @@ def __str__(self): | |||
|
|
|||
|
|
|||
| class Interfaces(TraceableRecord): | |||
| device = Devices | |||
There was a problem hiding this comment.
Can you put this in a separate PR, it doesn't look like a speed improvement but a bug fix (if a bug fix and there isn't an issue for it, please open an issue).
| @@ -53,8 +84,12 @@ def __init__(self, api, app, name, model=None): | |||
| endpoint=self.name, | |||
| ) | |||
| self._choices = None | |||
| self._init_cache() | |||
There was a problem hiding this comment.
This is only called in one place and has a function defined for it - can you just put it inline with a comment instead, not sure why we need a separate function for this?
| @@ -20,6 +20,37 @@ | |||
| RESERVED_KWARGS = () | |||
|
|
|||
|
|
|||
| class CachedRecordRegistry: | |||
There was a problem hiding this comment.
Maybe? Split this and the one in core/app.py into two separate PR's? Although both are caching - the seem like functionally separate functionality that can be reviewed separately (and potentially backed out separately if any issues).
|
@srfwx , can you address/answer Arthur's concerns/questions? Otherwise, we'll need to close this out as stale work before too long. Also, please address the conflict if it's not resolved by other changes. |
|
Hi @jnovinger , |
Fixes: #658 #625
Various refactoring allows for a substantial increase in performance of the Record initialization.
_parse_valueswas heavily refactored to avoid multiple similar checks and unnecessary processingget_returnnow only needs to work onRecordobject and drops legacy code from Netbox 2.7_endpoint_from_urlcan work with string instead of using theurlsplitlibrary