New ingestion_schemas + MANY minor fixes and improvements
#438
Conversation
…Centre/gl-issue-418 Allow reading pose model metadata from local folder
…o datajoint_pipeline
why? corrupted data in harp files? not sure
ttngu207
changed the title
ingestion_schemas + MANY minor fixes and improvements
aeon/dj_pipeline/__init__.py
Outdated
| Args: | ||
| query (datajoint.Query): A query object containing data from a Stream table | ||
| drop_pk (bool, optional): Drop primary key columns. Defaults to True. | ||
| round_microseconds (bool, optional): Round timestamps to microseconds. Defaults to False. |
There was a problem hiding this comment.
Mismatch is in the description of round_microseconds which says "Defaults to False"
| # Patch data - TriggerPellet, DepletionState, Encoder (distancetravelled) | ||
| # For wheel data, downsample to 10Hz | ||
| final_encoder_fs = 10 | ||
| final_encoder_hz = 10 |
There was a problem hiding this comment.
I think we actually want this to be 50 hz, not 10 hz
There was a problem hiding this comment.
The Encoder data is ingested at 50Hz, this is at the BlockAnalysis step, do we also want 50Hz here (or further downsample to 10?)
| if encoder_df.empty: | ||
| encoder_df["distance_travelled"] = 0 | ||
| else: | ||
| encoder_df["distance_travelled"] = -1 * analysis_utils.distancetravelled(encoder_df.angle) |
There was a problem hiding this comment.
maybe add a comment saying something like -1 is for placement of magnetic encoder, where wheel movement actually decreases encoder value?
| patch_rate = depletion_state_df.rate.iloc[0] | ||
| patch_offset = depletion_state_df.offset.iloc[0] | ||
| # handles patch rate value being INF | ||
| patch_rate = 999999999 if np.isinf(patch_rate) else patch_rate |
There was a problem hiding this comment.
is it actually an issue if patch rate is inf? Does it cause some downstream issue? We do this as default when no env is loaded.
There was a problem hiding this comment.
Yes, it's due to MySQL float doesn't handle INF well - we can convert to NaN but that would lose the intended meaning of INF here
| @@ -288,27 +299,50 @@ def make(self, key): | |||
| & f'chunk_start <= "{chunk_keys[-1]["chunk_start"]}"' | |||
| )[:block_start] | |||
| subject_visits_df = subject_visits_df[subject_visits_df.region == "Environment"] | |||
| subject_visits_df = subject_visits_df[~subject_visits_df.id.str.contains("Test", case=False)] | |||
There was a problem hiding this comment.
sometimes we use other, non "Test" subjects as test subjects. Maybe the check should be, if the subject does not begin with 'baa' (can str.lower to check for regardless of case) ?
There was a problem hiding this comment.
I see. Is startswith("baa") reliable, future-proof? Perhaps better logic is to also cross check with the Subjects to be manually specified by users for a particular experiment
| ) | ||
| pos_df = fetch_stream(pos_query)[block_start:block_end] | ||
| pos_df["likelihood"] = np.nan | ||
| # keep only rows with area between 0 and 1000 |
There was a problem hiding this comment.
is this because areas of > 1000 is likely an experimenter, or some other artifact? Maybe specify that in the comment?
| data = pd.DataFrame(payload, index=seconds, columns=self.columns) | ||
|
|
||
| # remove rows where the index is zero (why? corrupted data in harp files?) | ||
| data = data[data.index != 0] |
There was a problem hiding this comment.
This I guess will be fixed in the next PR targeting a new Encoder reader in ingestion_schemas.py ?
| @@ -48,6 +48,12 @@ def __init__(self, path): | |||
| super().__init__(_reader.Pose(f"{path}_test-node1*")) | |||
|
|
|||
|
|
|||
| class Pose03(Stream): | |||
There was a problem hiding this comment.
made add a comment that this is necessary due to changing registers for the pose streams for social02 in particular? And that 03 corresponds to the fact that this is because this pattern is what we're going with for social03 and moving forward? Or call this class something else?
There was a problem hiding this comment.
Or just import Pose from social_03.py as Pose03? I think I like something like this actually
There was a problem hiding this comment.
This is primarily due to full-pose data for social0.2 has a different pattern than that of the non-fullpose SLEAP data.
I don't think you can do from social_03 import Pose as Pose03, the way the the Devices/Streams being instantiated is that it's using the class name, so aliasing wouldn't work.
There was a problem hiding this comment.
I've left a bunch of small comments here.
A general point - maybe add top-level module docstrings for all these modules? Unless this will be done by @MilagrosMarin in #443 ?
Yes, this will be address when we apply Ruff checks and fixes |
This PR includes a few major new logic/features and many minor fixes (mostly for BlockAnalysis)
ingestion_schemas- i.e. a separate set of schemas for DJ ingestion (specializedEncoderreader andVideoreader)hw_timestampcolumnstreams.pyDJ schema accordingly (mainly to drop extra columns when reading Video data)social0.2BlobPositionto read and store blob position tracking when SLEAP data is not availablesubject_in_patchtimesfetch_streamfunction rounds times to microseconds (mysql precision is to microseconds only)BlobPositionwhenSLEAPTrackingis not availableBlockDetection- when double 0s are found, use the first 0s (instead of the 2nd one)BlockForagingcomputed tableFix #427
Replace #437