vidjil issueshttps://gitlab.inria.fr/vidjil/vidjil/-/issues2023-11-06T15:08:58+01:00https://gitlab.inria.fr/vidjil/vidjil/-/issues/5188Admin group not displayed in patients list2023-11-06T15:08:58+01:00CHESNIN ClementAdmin group not displayed in patients listWhen displaying the `patients` list view, if the groups contains `admin` it is not displayed (cf Cypress test_user 03)
When investigating a bit in the code, it seems this was done on purpose (see `get_groups_string` in `SampleSet.py`)....When displaying the `patients` list view, if the groups contains `admin` it is not displayed (cf Cypress test_user 03)
When investigating a bit in the code, it seems this was done on purpose (see `get_groups_string` in `SampleSet.py`). We talked about this with @fthonier, but he does not know either what this was done.
Is this the expected behavior ?https://gitlab.inria.fr/vidjil/vidjil/-/issues/5186py4web; fix review server2024-01-22T14:13:34+01:00THONIER Florianpy4web; fix review serverSince conversion to py4web; job to start a review server don't work.
Job success, but opening link return to error nginx 404.
Proxy, conf.js, docker env ?Since conversion to py4web; job to start a review server don't work.
Job success, but opening link return to error nginx 404.
Proxy, conf.js, docker env ?Web 2024.04https://gitlab.inria.fr/vidjil/vidjil/-/issues/5183Py4web; update documentation2024-01-22T14:14:19+01:00THONIER FlorianPy4web; update documentationSome documentation need to be added and updated, on dev, usage, migration, ...Some documentation need to be added and updated, on dev, usage, migration, ...Web 2024.04https://gitlab.inria.fr/vidjil/vidjil/-/issues/5181py4web; Is script create_clone_db_py working well ?2024-01-22T15:39:10+01:00THONIER Florianpy4web; Is script create_clone_db_py working well ?Make some tests to know if everything work well with new backend server.Make some tests to know if everything work well with new backend server.Web 2024.04https://gitlab.inria.fr/vidjil/vidjil/-/issues/5177Top by locus, refine criteria2023-10-20T11:57:05+02:00Mathieu GiraudTop by locus, refine criteriaThe following discussion from !1335 should be addressed:
* [ ] @mikael-s started a [discussion](https://gitlab.inria.fr/vidjil/vidjil/-/merge_requests/1335#note_852548 "algo: Add top by locus"): (+1 comment)
> There is no option to m...The following discussion from !1335 should be addressed:
* [ ] @mikael-s started a [discussion](https://gitlab.inria.fr/vidjil/vidjil/-/merge_requests/1335#note_852548 "algo: Add top by locus"): (+1 comment)
> There is no option to modify it (yet).
>
> Even the criterion itself is questionable: we are sure that we don't want to actually get the top 10 clones from **all** loci. However how can we determine the limit? What is a good enough locus?https://gitlab.inria.fr/vidjil/vidjil/-/issues/5176Similarity between clones; Add shortcut to select similar clonotype.2023-10-19T14:51:18+02:00THONIER FlorianSimilarity between clones; Add shortcut to select similar clonotype.Thinking about shourtcut on similarity:
* Shift+click on a clonotype: Select clone + similar clonotype. Doable after #5175
* Shift+ctrl+click on a clonotype: Select all clone with same V/J ?Thinking about shourtcut on similarity:
* Shift+click on a clonotype: Select clone + similar clonotype. Doable after #5175
* Shift+ctrl+click on a clonotype: Select all clone with same V/J ?https://gitlab.inria.fr/vidjil/vidjil/-/issues/5175Warning similarity; Position not correct after a fuse2023-10-19T14:51:00+02:00THONIER FlorianWarning similarity; Position not correct after a fuseFor the moment, we throw some warning on similarities between clones. (Similar to clone "\#1 - TRGV9*01 0/AC/1 TRGJ1*02").
After a fuse, clone \#1 is not correct. We need to update this position at the fuse.
In fact, even without fuse w...For the moment, we throw some warning on similarities between clones. (Similar to clone "\#1 - TRGV9*01 0/AC/1 TRGJ1*02").
After a fuse, clone \#1 is not correct. We need to update this position at the fuse.
In fact, even without fuse with other sample, position is false (see https://app.vidjil.org/3296-25?clone=42,64).https://gitlab.inria.fr/vidjil/vidjil/-/issues/5174Show N in sequence with a color2023-10-18T14:51:01+02:00THONIER FlorianShow N in sequence with a colorSome sequences have a N. We could probably give them a specific color to easily identify them in the interface.Some sequences have a N. We could probably give them a specific color to easily identify them in the interface.THONIER FlorianTHONIER Florianhttps://gitlab.inria.fr/vidjil/vidjil/-/issues/5173Merge clonotype on exact VDJ sequence (excluding primers)2023-10-19T14:34:09+02:00THONIER FlorianMerge clonotype on exact VDJ sequence (excluding primers)In some case, user upload data with primers. We have in these cases mix of primer of various length able to be used for a same clonotype. And in some configuration (no-clustering), we saw 2 clonotypes or more that are an artefact.
As p...In some case, user upload data with primers. We have in these cases mix of primer of various length able to be used for a same clonotype. And in some configuration (no-clustering), we saw 2 clonotypes or more that are an artefact.
As position and size of primer can vary, we can't make only a trimming of it.
Can we make a merge based on sequence but limited to Vstart ?
See case here: [59910-50?clone=27,35,45,81](https://app.vidjil.org/59910-50?clone=27,35,45,81)https://gitlab.inria.fr/vidjil/vidjil/-/issues/5172consensus sequence length: mix of sequence with a nucleotide mix at a positio...2023-10-18T17:56:11+02:00THONIER Florianconsensus sequence length: mix of sequence with a nucleotide mix at a position in the V geneSee sample 59910-50, with either configurations [IGH](https://app.vidjil.org/59910-2?clone=26) or [no-clusterign](https://app.vidjil.org/59910-50?).
We can see that in multi, top clonotype have a short consensus length for read that are...See sample 59910-50, with either configurations [IGH](https://app.vidjil.org/59910-2?clone=26) or [no-clusterign](https://app.vidjil.org/59910-50?).
We can see that in multi, top clonotype have a short consensus length for read that are able to get entire V gene. We can see in no clustering that top 2 clonotypes have a mix of C/T at this specific position that cut the consensus sequence.
In these case, maybe we can extend consensus but show (but how?) the mix of nucleotides.
* bypass mix with a color scale to represent variation. Can also be used to show variation in case of merge of clonotype.
* hover can show precise values variationhttps://gitlab.inria.fr/vidjil/vidjil/-/issues/5168Build our docker images on Alpine rather than Ubuntu2024-03-27T14:32:22+01:00Mikaël SalsonBuild our docker images on Alpine rather than UbuntuImages should be lighter.
Discussed with @clement.chesnin @fthonierImages should be lighter.
Discussed with @clement.chesnin @fthonierhttps://gitlab.inria.fr/vidjil/vidjil/-/issues/5167Can create generic set with some rules to select samples to import2023-10-11T18:14:43+02:00THONIER FlorianCan create generic set with some rules to select samples to importI want for a meta-analysis create a set with some samples that have in my case a similar tags.
That make me think about a dynamic playlist for music, with a list of rules to compose it.
It could be usefull to be able de to create a set...I want for a meta-analysis create a set with some samples that have in my case a similar tags.
That make me think about a dynamic playlist for music, with a list of rules to compose it.
It could be usefull to be able de to create a set by giving some rules, for example tags, number of reads, age of patient, disease, ...).
In this case, sample of a set should probably be constant with a refresh button if needed to update list.
This would improve usage of server for research purpose I presume.
This usage could also be benifit for appstats.
A button to generate a fuse without needing of relaunch an analysis should also be present I think.https://gitlab.inria.fr/vidjil/vidjil/-/issues/5164Non-recombined sequence leading to IGK false positive.2023-09-07T16:31:35+02:00Mikaël SalsonNon-recombined sequence leading to IGK false positive.See ~"NAN - Nantes" email on 2023-08-10 mentioning this issue:
He points this sequence:
```
ATGTGGACTCCCTCATGAGCAGATGCCACCAGGGCCACTGGCCCCAGCTTCCTCCTTCACAGCTGCAGTGGGGGCTGGGGCTGGGGCATCCCAGGGAGGGTTTTTGTATGAGCCTGTGTCACAGTGTGTGGTATTCGGCGGAG...See ~"NAN - Nantes" email on 2023-08-10 mentioning this issue:
He points this sequence:
```
ATGTGGACTCCCTCATGAGCAGATGCCACCAGGGCCACTGGCCCCAGCTTCCTCCTTCACAGCTGCAGTGGGGGCTGGGGCTGGGGCATCCCAGGGAGGGTTTTTGTATGAGCCTGTGTCACAGTGTGTGGTATTCGGCGGAGGGACCAAGCTGACCGTCCTAGGTGAGTCTCTTCTCCCCTCTCCTTCCCCGCTCTTGGGACAATTTCTGCTGTTTTTGTTTGTTTCTGTATCTTGTCTCA
```
where we find a `IGKV2D-29 0/78/7 J4`https://gitlab.inria.fr/vidjil/vidjil/-/issues/5162Contrôleur pour supprimer les données associées à un compte2024-01-18T14:50:43+01:00Mikaël SalsonContrôleur pour supprimer les données associées à un compteIl faudrait pouvoir supprimer toutes les données associées à un compte.
À faire avec prudence néanmoins quand l'utilisateur appartient à un groupe commun.
On se pose la question du RGPD : les données uploadées et analysées sont-elles de...Il faudrait pouvoir supprimer toutes les données associées à un compte.
À faire avec prudence néanmoins quand l'utilisateur appartient à un groupe commun.
On se pose la question du RGPD : les données uploadées et analysées sont-elles des données personnelles et soumises au RGPD (ie. si un utilisateur nous demande à tout supprimer de ce qu'il a uploadé est-on obligé de le faire, a fortiori quand il est dans un groupe) ?https://gitlab.inria.fr/vidjil/vidjil/-/issues/5161Analysis with long reads (thousands of pb)2023-08-30T10:24:56+02:00THONIER FlorianAnalysis with long reads (thousands of pb)We have a user that load data with long reads (thousands of pb).
Results are here and are not very usefull : https://app.vidjil.org/59029-25?
We saw 20% of VDJ. Majority of "Possibly XXX".We have a user that load data with long reads (thousands of pb).
Results are here and are not very usefull : https://app.vidjil.org/59029-25?
We saw 20% of VDJ. Majority of "Possibly XXX".https://gitlab.inria.fr/vidjil/vidjil/-/issues/5160Py4Web: changement de comportement de VidjilAuth2023-08-31T16:09:25+02:00Mikaël SalsonPy4Web: changement de comportement de VidjilAuthOn a vu avec @fthonier que VidjilAuth est désormais défini une fois pour toute dans Py4Web, alors qu'auparavant il était redéfini à chaque requête.
On stockait dans VidjilAuth des permissions pour ne pas avoir à les recharger plein de f...On a vu avec @fthonier que VidjilAuth est désormais défini une fois pour toute dans Py4Web, alors qu'auparavant il était redéfini à chaque requête.
On stockait dans VidjilAuth des permissions pour ne pas avoir à les recharger plein de fois lors d'un même appel à un contrôleur. Ce n'est plus possible puisque VidjilAuth sert désormais à la fois pour tous les utilisateurs.
Il faudra donc virer `self.permissions` dans `VidjilAuth` et voir l'impact en ressources.https://gitlab.inria.fr/vidjil/vidjil/-/issues/5159Ł caracter create a failure on server2023-08-16T11:35:39+02:00THONIER FlorianŁ caracter create a failure on serverThis character (Ł)[https://fr.wikipedia.org/wiki/%C5%81] create a fail at the preprocess step.
It is include by (unicode)[https://www.unicodepedia.com/unicode/latin-extended-a/142/latin-small-letter-l-with-stroke/].
In that case, no resu...This character (Ł)[https://fr.wikipedia.org/wiki/%C5%81] create a fail at the preprocess step.
It is include by (unicode)[https://www.unicodepedia.com/unicode/latin-extended-a/142/latin-small-letter-l-with-stroke/].
In that case, no result are returned, no file created, no log neither.
If I manually launch preprocess, it seem to complete normally. But the file name replace this character by octal utf encoding ('\305\202')
```bash
python flash2.py /usr/share/pear/ /mnt/upload/uploads/sequence_file.data_file.xxx.gz /mnt/upload/uploads/sequence_file.data_file.YYY.gz /mnt/result/tmp/pre/out-119703//AAA_łxxx.fastq.gz -r2 -f "-M 300"
--> created file: AAA_'$'\305\202''xxx.fastq.gz
```
I presume that error are with the server.
This issue concern web2py version of server. Don't know for futur py4web version.https://gitlab.inria.fr/vidjil/vidjil/-/issues/5158Non-recombined sequences beyond D7-27/J12023-10-06T11:09:00+02:00Mathieu GiraudNon-recombined sequences beyond D7-27/J1
As the sequencing lengths increase, some user reported us the following sequences that may be non-recombined as "IGHD7-27 - IGHJ1" (see #2232/!242)
- IGLV4-60 - IGLJ5
- TRBV6-9 - TRBJ2-2P
- TRBV6-1 - TRBJ2-2P
Todo !1345:
- [x] extend ...
As the sequencing lengths increase, some user reported us the following sequences that may be non-recombined as "IGHD7-27 - IGHJ1" (see #2232/!242)
- IGLV4-60 - IGLJ5
- TRBV6-9 - TRBJ2-2P
- TRBV6-1 - TRBJ2-2P
Todo !1345:
- [x] extend segment.cpp to handle other cases
- [ ] check these three cases, make tests such as `data/D7-27--J1.fa`
- [ ] fill the cases in segment.hhttps://gitlab.inria.fr/vidjil/vidjil/-/issues/5156Get some metrics from a server2024-01-22T14:22:32+01:00THONIER FlorianGet some metrics from a serverTHinking about it with @mikael-s : A way to get some metrics is to use prometheus and grafana. And as by default Prometheus use web resquest to get metrics, a way to get it with security is to use vidjil API to ask, as admin. APi seem to...THinking about it with @mikael-s : A way to get some metrics is to use prometheus and grafana. And as by default Prometheus use web resquest to get metrics, a way to get it with security is to use vidjil API to ask, as admin. APi seem to be a good way for that.
I try to create a really small dockerized flask application that will execute API to ask metrics and serve them locally.
My quick overview failed since the API is incompatible with py4web (other issue/MR).
Is it a good way to do it ?Web 2024.04https://gitlab.inria.fr/vidjil/vidjil/-/issues/5153Envoyer les données vers vdj.online2023-07-13T09:22:10+02:00Mikaël SalsonEnvoyer les données vers vdj.onlinehttps://vdj.online/
Service de MiXCR.
À discuter, peut-être pas très pratique à intégrer, car il faut renseigner pas mal d'info sur le protocole expérimental.https://vdj.online/
Service de MiXCR.
À discuter, peut-être pas très pratique à intégrer, car il faut renseigner pas mal d'info sur le protocole expérimental.