Add functions to convert from met to ET and UTC #1210
Conversation
subagonsouth
requested review from
nkerman,
tech3371,
laspsandoval and
torimarbois
| @@ -13,6 +13,30 @@ | |||
|
|
|||
| TICK_DURATION = 2e-5 # 20 microseconds as defined in imap_sclk_0000.tsc | |||
|
|
|||
| # Hard code the J2000 epoch. This allows for CDF epoch to be converted without | |||
| # use of SPICE though it should be noted that this results in a 5-second error | |||
| # due to the occurrence of 5 leap-seconds since the J2000 epoch. | |||
There was a problem hiding this comment.
Arguably a TODO should go here, since this is where the epoch is defined, and where the 5 leapsecond fix will go, right?
There was a problem hiding this comment.
True Nat. I meant to ask Tim if there is a ticket to revisit for this?
There was a problem hiding this comment.
I will add the TODO. Yes, there is a ticket for this. There are two ways to fix this each with a pro and con. See ticket: #1208
To clarify, this J2000EPOCH value is correct. The bug is in the code in the cdf/utils.py here:
dt64 = J2000_EPOCH + dataset["epoch"].values[0].astype("timedelta64[ns]")
start_time = np.datetime_as_string(dt64, unit="D").replace("-", "")
| # use of SPICE though it should be noted that this results in a 5-second error | ||
| # due to the occurrence of 5 leap-seconds since the J2000 epoch. | ||
| # The UTC string was generated by: | ||
| # >>> spiceypy.et2utc(0, "ISOC", 9) |
There was a problem hiding this comment.
Thanks for these doc strings. Helps someone like me.
imap_processing/spice/time.py
Outdated
| Returns | ||
| ------- | ||
| numpy.ndarray[str] | ||
| The mission elapsed time converted to UTC string. |
There was a problem hiding this comment.
Can you add an example about the format it returns in? If I understand correctly, ISOC format is something like YYYY-MM-DDTHH:MM:SSZ. Is that right? or is it like this 1987-04-12T16:31:12.814?
| """ | ||
| if isinstance(met, typing.Iterable): | ||
| return np.asarray([np.datetime64(utc) for utc in met_to_utc(met)]) | ||
| return np.datetime64(met_to_utc(met)) |
There was a problem hiding this comment.
These two functions are gonna be so helpful! Thank you for adding those
| dt64 = met_to_datetime64(met) | ||
| np.testing.assert_array_equal( | ||
| dt64.astype("datetime64[us]"), expected_dt64.astype("datetime64[us]") | ||
| ) |
No, actually, it is a bit more subtle. That epoch definition is correct in that it is the number of nanoseconds elapsed since the Unix epoch (00:00:00 UTC on 1 January 1970). The problem is that |
subagonsouth
merged commit aa2ecc4
into
IMAP-Science-Operations-Center:dev
…ons-Center#1210) * Add functions to convert from met to ET and UTC * Switch MAG and GLOWS imports to use spice.time definition of J2000_EPOCH * PR feedback
Change Summary
Overview
Adds functions for converting from Mission Elapsed Time (MET) to ET and UTC
Updated Files
J2000_EPOCHmoved into the spice.time.py moduleJ2000_EPOCHmoved into the spice.time.py modulemet_to_sclkticksfunction for reusemet_to_utcfunctionmet_to_datetime64function with SPICE backed implementation.spice/time.pyfunctions.spiceypy as spicebecause it is confusing.Closes: #1183