Issue: UDA URI fails with cache_mode=none mode

[prasad-sawantdesai]

When opening the following UDA URI, it fails with a "source not found" exception.

Environment

imas-core==5.7.0
imas-data-dictionaries==4.1.1
imas-python==2.2.0

import imas

entry = imas.DBEntry(
    "imas://uda.iter.org/uda?path=/work/imas/shared/imasdb/ITER/3/134173/106&backend=hdf5&cache_mode=none",
    "r",
)
summary = entry.get("summary", autoconvert=False)

Python 3.11.5 (main, Jul 30 2025, 19:17:56) [GCC 13.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import imas
>>> t= imas.DBEntry("imas://uda.iter.org/uda?path=/work/imas/shared/imasdb/ITER/3/134173/106&backend=hdf5&cache_mode=none", "r")
09:54:57 INFO     Parsing data dictionary version 4.1.1 @dd_zip.py:89
>>> sum = t.get("summary", autoconvert=False)
09:55:09 INFO     Parsing data dictionary version 3.38.1 @dd_zip.py:89
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/ITER/sawantp1/github/SimDB/myenv/lib/python3.11/site-packages/imas/db_entry.py", line 356, in get
    return self._get(
           ^^^^^^^^^^
  File "/home/ITER/sawantp1/github/SimDB/myenv/lib/python3.11/site-packages/imas/db_entry.py", line 622, in _get
    return self._dbe_impl.get(
           ^^^^^^^^^^^^^^^^^^^
  File "/home/ITER/sawantp1/github/SimDB/myenv/lib/python3.11/site-packages/imas/backends/imas_core/db_entry_al.py", line 232, in get
    get_children(destination, read_ctx, time_mode, nbc_map)
  File "/home/ITER/sawantp1/github/SimDB/myenv/lib/python3.11/site-packages/imas/backends/imas_core/db_entry_helpers.py", line 65, in get_children
    get_children(element, ctx, time_mode, nbc_map)
  File "/home/ITER/sawantp1/github/SimDB/myenv/lib/python3.11/site-packages/imas/backends/imas_core/db_entry_helpers.py", line 69, in get_children
    data = ctx.read_data(new_path, timebase, data_type.al_type, ndim)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/ITER/sawantp1/github/SimDB/myenv/lib/python3.11/site-packages/imas/backends/imas_core/al_context.py", line 149, in read_data
    status, data = ll_interface.read_data(
                   ^^^^^^^^^^^^^^^^^^^^^^^
  File "imas_core/_al_lowlevel.pyx", line 992, in imas_core._al_lowlevel.al_read_data
  File "imas_core/_al_lowlevel.pyx", line 975, in imas_core._al_lowlevel.al_read_data_string
imas_core.exception.ALException: Error while reading data: al_plugin_read_data: cannot find node summary/ids_properties/source in data dictionary (source not found)
Error status=-1

[maarten-ic]

Hi Prasad,

This is a limitation of the UDA backend with different data dictionary versions. See also the documentation on this: https://imas-python.readthedocs.io/en/stable/multi-dd.html#uda-backend-caching-and-data-dictionary-versions.
You could use lazy loading in addition to disabling the UDA cache, or use one of the two other suggestions from the documentation to work around this problem.

[olivhoenen]

The problem is that with this data, the second workaround option (cache_mode=none) does not seem to work, and lazy loading does not change this, as soon as you try reading ids_properties you end up with this error.

Either this second option does not really work and we should remove it from the doc, or it has a bug (maybe in IMAS-Python, but could also be in UDA backend of IMAS-Core or in UDA-Plugins) and can be fixed, or it has known limitation that we faced in this particular dataset (and we could add a note in the doc about it).

[maarten-ic]

The problem is that with this data, the second workaround option (cache_mode=none) does not seem to work, and lazy loading does not change this, as soon as you try reading ids_properties you end up with this error.

Ah right, I didn't test before posting my previous response 😅.

The following does seem to work from SDCC:

>>> import imas
>>> entry = imas.DBEntry(
...     "imas://uda.iter.org/uda?path=/work/imas/shared/imasdb/ITER/3/134173/106&backend=hdf5&cache_mode=none",
...     "r",
... )
>>> summary = entry.get("summary", autoconvert=False, lazy=True)
>>> summary.ids_properties
<IDSStructure (IDS:summary, ids_properties)>
>>> summary.ids_properties.homogeneous_time
<IDSInt0D (IDS:summary, ids_properties/homogeneous_time, INT_0D)>
int(1)
>>> summary.ids_properties.version_put.data_dictionary
<IDSString0D (IDS:summary, ids_properties/version_put/data_dictionary, STR_0D)>
str('3.38.1')

But indeed, when requesting summary.ids_properties.source the error is raised:

>>> summary.ids_properties.source
Traceback (most recent call last):
[...]
imas_core.exception.ALException: Error while reading data: al_plugin_read_data: cannot find node summary/ids_properties/source in data dictionary (source not found)
Error status=-1

Either this second option does not really work and we should remove it from the doc, or it has a bug (maybe in IMAS-Python, but could also be in UDA backend of IMAS-Core or in UDA-Plugins) and can be fixed, or it has known limitation that we faced in this particular dataset (and we could add a note in the doc about it).

Indeed.. the second option only seems to work when requesting data that exists both in the "default" and on-disk Data Dictionary version.

• I think there's no easy way to fix this:
  • The UDA backend seems to require that the node exists in its loaded IDSDef.xml (link to code), so that seems tricky to patch.
  • The other option I can think of is for IMAS-Python to detect that the source data is in a different DD version, and then automatically construct another DBEntry object in the background with that DD version (pretty much the third option from the docs). This would add complexity with automatic DD version conversions and resource management (especially when lazy loading, because we would need to track when the background DBEntry needs to stay alive and when it may be deleted safely). So this option may not be worth it.
• I'm okay with removing the second option from the documentation. Let me know if you want me to pick that up.

[olivhoenen]

Ok thanks @maarten-ic, let's just remove the 2nd workaround, cf #127