PLP package generated by WebAPI v2.12.0 fails on renv

[konstjar]

Expected behavior

As a user I'm able to run generated package from RStudio

Actual behavior

a. Wrong command in STUDY-PACKAGE-SETUP.md

# Download the lock file:
download.file("https://raw.githubusercontent.com/ohdsi-studies/regressPLP/main/renv.lock", "renv.lock")

b. The 'renv::init()' failes on the error:

Retrieving 'https://cloud.r-project.org/src/contrib/Archive/CohortGenerator/CohortGenerator_0.3.0.tar.gz' ...
Warning: failed to find binary for 'CohortGenerator 0.3.0' in package repositories
Warning: failed to find source for 'CohortGenerator 0.3.0' in package repositories
Warning: failed to retrieve 'https://cloud.r-project.org/src/contrib/Archive/CohortGenerator/CohortGenerator_0.3.0.tar.gz' [error code 22]
Error: failed to retrieve package 'CohortGenerator@0.3.0'
In addition: Warning message:
curl: (22) The requested URL returned error: 404 

Steps to reproduce behavior

1. Download zip file for PLP package from ATLAS UI
2. Follow steps in STUDY-PACKAGE-SETUP.md instruction
3. Run 'renv::init()' command

[konstjar]

More questions. CohortGenerator 0.3.0 fails. But package requires CohortGenerator >= 0.4.0 version. Does it mean renv.lock file is wrong?

[konstjar]

Also PLP package should not be higher v5.0.5? v6.0 is not compatible further because of some methods were changed

[anthonysena]

Hi @konstjar - yes, seems like there are a few problems here.

a. Wrong command in STUDY-PACKAGE-SETUP.md

Not sure what is wrong with this command? The presumption is that the package would be uploaded to the ohdsi-studies area for use.

b. The 'renv::init()' failes on the error....

The renv.lock file reference for CohortGenerator should point to GitHub. An example of how that should look is here: https://github.com/OHDSI/CohortGeneratorModule/blob/main/renv.lock#L420-L428. Next, we'll need to decide which version to use....

More questions. CohortGenerator 0.3.0 fails. But package requires CohortGenerator >= 0.4.0 version. Does it mean renv.lock file is wrong?

Presumably we should make use of CohortGenerator v0.4.0 and this should be the version referenced in the renv.lock file.

Also PLP package should not be higher v5.0.5? v6.0 is not compatible further because of some methods were changed

Yes, we should probably move that to v5.0.5 as v6 has breaking changes. We should also ensure this is referenced in the renv.lock file properly.

Appreciate you digging into this! As a suggestion here, could you make these changes locally and if there are any other issues encountered, can you post them here? Once your package is working, we can revise the skeleton reference in Hydra and plan another release.

[konstjar]

@anthonysena Thank you for quick response. I'm moving forward and collecting the set of changes to be addressed and my steps.

1. STUDY-PACKAGE-SETUP.md - I got it, it's intruction file for end users who download the analysis after it's published on ohdsi-studies. In this case I think STUDY-PACKAGE-BUILD.md instruction should be updated to reflect "renv". What do you think?

2. In reality, renv.lock file contains 2 CohortGenerator entries:

I removed first one with "CRAN".

3. I changed reference to CohortGenerator v0.4.0. After that I was able to restore environment with "renv".

I'm trying to run the analysis and I will update here on results

[konstjar]

I continue with exercise.

My initial platform was Macbook Pro with M1 processor. During restoring all packages I met the issue that RJSONIO package was not able to compile. I switched to Windows 10 platform. After setting up all things, I ran the analysis and got this error:

Creating cohorts
Connecting using PostgreSQL driver
Creating cohort tables
- Created table results_atlas_demo.plpanalysisCohort
- Created table results_atlas_demo.plpanalysisCohort_inclusion
- Created table results_atlas_demo.plpanalysisCohort_inclusion_result
- Created table results_atlas_demo.plpanalysisCohort_inclusion_stats
- Created table results_atlas_demo.plpanalysisCohort_summary_stats
- Created table results_atlas_demo.plpanalysisCohort_censor_stats
Creating cohort tables took 2.71secs
Loading cohortDefinitionSet
1/2- Generating cohort: Transfusion (4)                                                                                                                     
  |===============================================================================================================| 100%
Executing SQL took 1.34 mins
2/2- Generating cohort: ibuprofen (8)
  |===============================================================================================================| 100%
Executing SQL took 18.1 secs
Generating cohort set took 1.66 mins
Counting cohorts
Counting cohorts took 0.741 secs
Running predictions
Updating as cohort covariate settings is being used
Error in predictionAnalysisList$analyses[[i]] : subscript out of bounds
> 

Here is a sessionInfo:

> sessionInfo()
R version 4.1.2 (2021-11-01)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 22000)

Matrix products: default

locale:
[1] LC_COLLATE=English_Europe.1251  LC_CTYPE=English_Europe.1251    LC_MONETARY=English_Europe.1251
[4] LC_NUMERIC=C                    LC_TIME=English_Europe.1251    

attached base packages:
[1] stats     graphics  grDevices datasets  utils     methods   base     

other attached packages:
[1] plpanalysis_1.0.6

loaded via a namespace (and not attached):
 [1] Rcpp_1.0.7                   pillar_1.5.1                 compiler_4.1.2               DatabaseConnector_5.0.2     
 [5] tools_4.1.2                  bit_4.0.4                    jsonlite_1.7.2               lubridate_1.7.10            
 [9] checkmate_2.0.0              RSQLite_2.2.5                memoise_2.0.0                lifecycle_1.0.0             
[13] tibble_3.1.0                 pkgconfig_2.0.3              rlang_0.4.10                 rstudioapi_0.13             
[17] DBI_1.1.1                    parallel_4.1.2               PatientLevelPrediction_5.0.4 FeatureExtraction_3.1.1     
[21] fastmap_1.1.0                rJava_0.9-13                 withr_2.4.1                  CohortGenerator_0.4.0       
[25] dplyr_1.0.5                  generics_0.1.0               vctrs_0.3.7                  hms_1.0.0                   
[29] triebeard_0.3.0              bit64_4.0.5                  tidyselect_1.1.0             glue_1.6.1                  
[33] R6_2.5.0                     fansi_0.4.2                  vroom_1.5.7                  Andromeda_0.6.0             
[37] tzdb_0.1.1                   readr_2.1.2                  purrr_0.3.4                  ParallelLogger_2.0.1        
[41] SqlRender_1.7.0              blob_1.2.1                   magrittr_2.0.1               backports_1.2.1             
[45] urltools_1.7.3               ellipsis_0.3.1               assertthat_0.2.1             renv_0.13.2                 
[49] utf8_1.2.1                   stringi_1.7.6                cachem_1.0.4                 crayon_1.4.1                
> 

Next step will be to update PLP to v5.0.5

[konstjar]

Attempt to use PLP v5.0.5 failed with the same problem:

Creating cohorts
Connecting using PostgreSQL driver
Creating cohort tables
- Created table results_atlas_demo.plpanalysisCohort
- Created table results_atlas_demo.plpanalysisCohort_inclusion
- Created table results_atlas_demo.plpanalysisCohort_inclusion_result
- Created table results_atlas_demo.plpanalysisCohort_inclusion_stats
- Created table results_atlas_demo.plpanalysisCohort_summary_stats
- Created table results_atlas_demo.plpanalysisCohort_censor_stats
Creating cohort tables took 2.38secs
Loading cohortDefinitionSet
1/2- Generating cohort: Transfusion (4)                                                                                                                     
  |===============================================================================================================| 100%
Executing SQL took 19.8 secs
2/2- Generating cohort: ibuprofen (8)
  |===============================================================================================================| 100%
Executing SQL took 4.92 secs
Generating cohort set took 25.36 secs
Counting cohorts
Counting cohorts took 0.712 secs
Running predictions
Updating as cohort covariate settings is being used
Error in predictionAnalysisList$analyses[[i]] : subscript out of bounds
> 

Here is sessionInfo:

> sessionInfo()
R version 4.1.2 (2021-11-01)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 22000)

Matrix products: default

locale:
[1] LC_COLLATE=English_Europe.1251  LC_CTYPE=English_Europe.1251    LC_MONETARY=English_Europe.1251
[4] LC_NUMERIC=C                    LC_TIME=English_Europe.1251    

attached base packages:
[1] stats     graphics  grDevices datasets  utils     methods   base     

other attached packages:
[1] plpanalysis_1.0.6

loaded via a namespace (and not attached):
 [1] Rcpp_1.0.7                   compiler_4.1.2               pillar_1.5.1                 DatabaseConnector_5.0.2     
 [5] tools_4.1.2                  bit_4.0.4                    jsonlite_1.7.2               lubridate_1.7.10            
 [9] checkmate_2.0.0              RSQLite_2.2.5                memoise_2.0.0                lifecycle_1.0.0             
[13] tibble_3.1.0                 pkgconfig_2.0.3              rlang_0.4.10                 rstudioapi_0.13             
[17] DBI_1.1.1                    cli_3.2.0                    parallel_4.1.2               PatientLevelPrediction_5.0.5
[21] FeatureExtraction_3.2.0      fastmap_1.1.0                rJava_0.9-13                 withr_2.4.1                 
[25] CohortGenerator_0.4.0        dplyr_1.0.5                  generics_0.1.0               vctrs_0.3.7                 
[29] hms_1.0.0                    triebeard_0.3.0              bit64_4.0.5                  tidyselect_1.1.0            
[33] glue_1.6.1                   R6_2.5.0                     fansi_0.4.2                  vroom_1.5.7                 
[37] Andromeda_0.6.0              tzdb_0.1.1                   readr_2.1.2                  purrr_0.3.4                 
[41] ParallelLogger_3.0.1         SqlRender_1.7.0              blob_1.2.1                   magrittr_2.0.1              
[45] backports_1.2.1              urltools_1.7.3               ellipsis_0.3.1               assertthat_0.2.1            
[49] renv_0.13.2                  utf8_1.2.1                   stringi_1.7.6                cachem_1.0.4                
[53] crayon_1.4.1               

[konstjar]

I found similar issue in PLP package, but there is no solution shared:
OHDSI/PatientLevelPrediction#292

Problem happens here:
https://github.com/OHDSI/SkeletonPredictionStudy/blob/main/R/Main.R#L165

Do you have any suggestions for any next steps I can try?

[anthonysena]

Thanks @konstjar - can you post the JSON for the study if possible? I'd like to take a look at that as it would appear that the predictionAnalysisList$analyses[[i]] is going beyond the length of the list for some reason.

[konstjar]

Please find attached.
plp.txt

[iamalonso]

Hi @anthonysena, I'm facing the same error as @konstjar. The steps I follow was:

1. Create the Study Package with Hydra
2. Edit the Main.R as suggested in this thread

My JSON for the study is this and my sessionInfo():

R version 4.1.2 (2021-11-01)
Platform: aarch64-apple-darwin20 (64-bit)
Running under: macOS Big Sur 11.2.2
[plp.txt](https://github.com/OHDSI/WebAPI/files/10053386/plp.txt)

Matrix products: default
LAPACK: /Library/Frameworks/R.framework/Versions/4.1-arm64/Resources/lib/libRlapack.dylib

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] stats     graphics  grDevices datasets  utils     methods   base     

other attached packages:
[1] StudyClaveAZulPackage_1.0.6

loaded via a namespace (and not attached):
 [1] tidyselect_1.2.0             rJava_1.0-6                  vctrs_0.5.1                  generics_0.1.3               htmltools_0.5.3             
 [6] utf8_1.2.2                   blob_1.2.3                   rlang_1.0.6                  pillar_1.8.1                 later_1.3.0                 
[11] glue_1.6.2                   withr_2.5.0                  DBI_1.1.3                    ParallelLogger_3.0.1         bit64_4.0.5                 
[16] lifecycle_1.0.3              memoise_2.0.1                tzdb_0.3.0                   fastmap_1.1.0                httpuv_1.6.5                
[21] DatabaseConnector_5.1.0      fansi_1.0.3                  urltools_1.7.3               triebeard_0.3.0              Rcpp_1.0.9                  
[26] xtable_1.8-4                 readr_2.1.3                  renv_0.15.2                  backports_1.4.1              promises_1.2.0.1            
[31] checkmate_2.1.0              cachem_1.0.6                 vroom_1.6.0                  jsonlite_1.8.3               mime_0.12                   
[36] bit_4.0.5                    CohortGenerator_0.7.0        hms_1.1.2                    digest_0.6.30                stringi_1.7.8               
[41] dplyr_1.0.10                 shiny_1.7.1                  SqlRender_1.10.1             cli_3.4.1                    tools_4.1.2                 
[46] magrittr_2.0.3               FeatureExtraction_3.2.0      tibble_3.1.8                 RSQLite_2.2.18               crayon_1.5.2                
[51] pkgconfig_2.0.3              ellipsis_0.3.2               lubridate_1.8.0              assertthat_0.2.1             Andromeda_0.6.1             
[56] rstudioapi_0.14              R6_2.5.1                     PatientLevelPrediction_5.0.5 compiler_4.1.2              

[iamalonso]

Hello @konstjar, have you made any progress with this issue?

[konstjar]

@iamalonso Nothing yet, I hope to have some information after Thanksgiving.