Second gap found in the open-issue review with no issue tracking it. Companion
to #174.
records with an image 5/19
adopted image sources 6
Adopted and image-capable: wikimedia_commons, pmc_oa, emdb_empiar,
cell_image_library, bioimage_archive, cryoet_data_portal.
Imageless: type_iv_pilus, mreb_filament, divisome_complex, elongasome,
ATP synthase, cell_outer_membrane, peptidoglycan_based_cell_wall, gas_vesicle,
magnetosome, encapsulin_nanocompartment, nucleoid, chemoreceptor array,
injectisome, endospore.
A correction worth recording
I expected the cryoET Data Portal adoption to have closed two of these. It has
not, and cannot:
$ uv run python scripts/cryoet_data_portal.py
data/structures/inclusion/gas_vesicle.yaml dataset 10014 run 14004 unchanged
data/structures/appendage/type_iv_pilus.yaml dataset 10155 run 7978 ... unchanged
nothing to write
It matches exactly those two records and writes a datasets entry, not an
image — both still have images: 0. A dataset citation is real provenance and
worth having, but it puts no picture on the page.
This is the ADOPTED-cannot-express-coverage problem again: six sources are
adopted, and the question "does this record have a picture" is answered no for
14 of 19. A source being adopted says a script exists and its licence was
checked; it says nothing about where it has been applied, or whether the artifact
it produces is the one the gap needs.
Suggested next
pmc_oa is the one demonstrated to produce a hostable micrograph end to end
(#35, PMC9968351 Fig. 1, CC BY 4.0). Its coverage subcommand reports
figure-level candidates rather than article hits, so it can be pointed at a named
record:
uv run python scripts/pmc_oa.py coverage --term '<structure> AND (micrograph OR cryo-EM)'
Worth doing for the largest, most-imaged structures first — injectisome,
magnetosome, nucleoid — and recording a no_image_yet CURATION_TODO on those
where an openly licensed micrograph genuinely does not exist, as magnetosome
already does.
Belongs to the source-queue skill's territory for choosing which source; this
issue is the gap measurement, not the plan.
Second gap found in the open-issue review with no issue tracking it. Companion
to #174.
Adopted and image-capable:
wikimedia_commons,pmc_oa,emdb_empiar,cell_image_library,bioimage_archive,cryoet_data_portal.Imageless: type_iv_pilus, mreb_filament, divisome_complex, elongasome,
ATP synthase, cell_outer_membrane, peptidoglycan_based_cell_wall, gas_vesicle,
magnetosome, encapsulin_nanocompartment, nucleoid, chemoreceptor array,
injectisome, endospore.
A correction worth recording
I expected the cryoET Data Portal adoption to have closed two of these. It has
not, and cannot:
It matches exactly those two records and writes a
datasetsentry, not animage — both still have
images: 0. A dataset citation is real provenance andworth having, but it puts no picture on the page.
This is the
ADOPTED-cannot-express-coverage problem again: six sources areadopted, and the question "does this record have a picture" is answered no for
14 of 19. A source being adopted says a script exists and its licence was
checked; it says nothing about where it has been applied, or whether the artifact
it produces is the one the gap needs.
Suggested next
pmc_oais the one demonstrated to produce a hostable micrograph end to end(#35, PMC9968351 Fig. 1, CC BY 4.0). Its
coveragesubcommand reportsfigure-level candidates rather than article hits, so it can be pointed at a named
record:
Worth doing for the largest, most-imaged structures first — injectisome,
magnetosome, nucleoid — and recording a
no_image_yetCURATION_TODOon thosewhere an openly licensed micrograph genuinely does not exist, as
magnetosomealready does.
Belongs to the
source-queueskill's territory for choosing which source; thisissue is the gap measurement, not the plan.