Out of memory caching #25
Conversation
Add cache_reset
| BugReports: https://github.com/hadley/memoise/issues | ||
| Imports: | ||
| digest (>= 0.6.3) | ||
| digest (>= 0.6.3), | ||
| googleAuthR, |
There was a problem hiding this comment.
cache specific dependencies should probably go in Suggests: instead of Imports:
| cache_reset <- function() { | ||
| query_results <- query_ds(the_body = list(gqlQuery = list(queryString = paste0("SELECT * FROM ", cache_name)))) | ||
| while((query_results$batch$moreResults != "NO_MORE_RESULTS") | is.null(query_results$batch$entityResults) == FALSE) { | ||
| ids <- (query_results$batch$entityResults$entity$key$path %>% dplyr::bind_rows())$name |
There was a problem hiding this comment.
You will need to remove use of %>% and the dplyr dependency here.
I don't know exactly what data structure is being returned, but from the code this looks like you can use vapply() with an extractor function.
|
@danielecook Please look at the travis build results (https://travis-ci.org/hadley/memoise/builds/148021731#L1180-L1217) and fix all Warnings, most are due to documentation mismatches or unimported functions. |
|
@jimhester Hey thanks! I wasn't sure whether the authors would want to add this so I renamed the package and had started my own fork but I'll change it back and resolve these issues. I'll also add a file-based cache soon (for dropbox/Google Drive). I was in a bit of a rush to get started using this for some of my own projects (-; |
Current coverage is 36.11% (diff: 0.00%)@@ master #25 diff @@
==========================================
Files 2 5 +3
Lines 93 252 +159
Methods 0 0
Messages 0 0
Branches 0 0
==========================================
- Hits 92 91 -1
- Misses 1 161 +160
Partials 0 0
|
|
Got this to work, but had to remove quite a bit of documentation. I tend not to be so stringent as to not allow things like: |
|
They should be in #' \dontrun{
#' # ... code to not run
#' }See (http://r-pkgs.had.co.nz/man.html#man-functions) for a good reference. |
|
I ended up merging this with b3b9ae6. I cleaned up some things, but decided to remove the google datastore backend, there was too much API specific code that I don't think belongs in this package. Also I was a little confused why you were using datastore for this, Google cloud store (with https://github.com/cloudyr/googleCloudStorageR) seems like a better fit. Thank you again for the PR and sorry it took so long to get merged. I think this is a very nice improvement to the package! |
|
Hey just saw these options were folded in. Thanks for taking care of this! Google Cloud storage is a better fit but it's also probably slower. Regardless, glad to see these changes implemented. Have a good rest of the year! |
No description provided.