Vince Carey vjcitn
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| library(EBImage) | |
| library(shiny) | |
| ui = fluidPage( | |
| sidebarLayout( | |
| sidebarPanel( | |
| helpText("EBImage explorer"), | |
| fileInput("inimg", "file"), | |
| numericInput("scalefactor", "multfac", min=1, max=50, value=10,step=1), | |
| sliderInput("blursig", "sigma for blur", min=1, max=100, value=50) |
vjcitn
/ filter_gtex.txt
Created
August 10, 2024 11:48
script for filtering a gtex resource from https://ftp.ebi.ac.uk/pub/databases/spot/eQTL/imported/GTEx_V8/ge/
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| #tabix Lung.tsv.gz 1:1-300000000 > lung.chr1.tsv | |
| #tabix Lung.tsv.gz 2:1-300000000 > lung.chr2.tsv | |
| #tabix Lung.tsv.gz 3:1-300000000 > lung.chr3.tsv | |
| #tabix Lung.tsv.gz 4:1-300000000 > lung.chr4.tsv | |
| #tabix Lung.tsv.gz 5:1-300000000 > lung.chr5.tsv | |
| #tabix Lung.tsv.gz 6:1-300000000 > lung.chr6.tsv | |
| #tabix Lung.tsv.gz 7:1-300000000 > lung.chr7.tsv | |
| #tabix Lung.tsv.gz 8:1-300000000 > lung.chr8.tsv | |
| #tabix Lung.tsv.gz 9:1-300000000 > lung.chr9.tsv | |
| #tabix Lung.tsv.gz 10:1-300000000 > lung.chr10.tsv |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| # retrieve these from https://mghp.osn.xsede.org/bir190004-bucket01/BiocXenData/ | |
| #-rw-r--r-- 1 exouser exouser 3300521324 Mar 20 21:17 transcripts.parquet | |
| #-rw-r--r-- 1 exouser exouser 68454210 Mar 20 21:16 nucleus_boundaries.parquet | |
| #-rw-r--r-- 1 exouser exouser 73791358 Mar 20 21:15 cell_boundaries.parquet | |
| #-rw-r--r-- 1 exouser exouser 242459483 Apr 10 03:46 cell_feature_matrix.tar.gz | |
| #-rw-r--r-- 1 exouser exouser 44907408 Mar 20 21:15 cells.csv.gz | |
| # tar zxf cell_feature_matrix.tar.gz to obtain folder cell_feature_matrix | |
| library(Matrix) |
vjcitn
/ explore_bert.R
Last active
July 17, 2024 13:35
look at 'safetensors' component of bert (use after usebert.R succeeds)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| # probe R user cache file system for bert resources | |
| explore_bert = function() { | |
| if (!requireNamespace("tibble")) stop("please install tibble to use this function") | |
| res = dir(tools::R_user_dir("../huggingface/hub/models--google-bert--bert-base-uncased/snapshots/", "cache"), full=TRUE, recursive=TRUE) | |
| if (length(res)<1) stop("not finding huggingface/hub or bert components in cache") | |
| safetpath = grep("model.safetensors", res, value=TRUE) | |
| st = try(reticulate::import("safetensors")) | |
| if (inherits(st, "try-error")) stop("can't import safetensors with reticulate") | |
| oo = st$safe_open(filename=safetpath, framework="pt") | |
| kk = oo$keys() |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| use_bert = function(phrase) { | |
| # use reticulate::py_install(c("torch", "transformers"), pip=TRUE) to set up | |
| # devtools::source_gist() may produce some warnings related to GPU | |
| # note that first run will populate .cache/huggingface/hub with model components | |
| my_bert_template = " | |
| # ensure accessible python has transformers installed | |
| from transformers import AutoTokenizer, BertForMaskedLM, logging | |
| from transformers import pipeline | |
| logging.set_verbosity_error() |
vjcitn
/ get68proc.R
Last active
June 16, 2024 15:38
function that will create a SingleCellExperiment by retrieving a serialized version of 'processed' PBMC 68k from Open Storage Network
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| get68proc = function(cache=BiocFileCache::BiocFileCache(), | |
| targetfolder=tempdir()) { | |
| zippath = "https://mghp.osn.xsede.org/bir190004-bucket01/BiocMatrixGenerics/pbmc68kproc.zip" | |
| ent = BiocFileCache::bfcquery(cache, "pbmc68kproc.zip") | |
| if (nrow(ent)==0) { | |
| ent = BiocFileCache::bfcadd(cache, rname=zippath, action="copy") | |
| } | |
| refresh = BiocFileCache::bfcquery(cache, "pbmc68kproc.zip") | |
| nzip = nrow(refresh) | |
| ind = 1 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| library(TENxPBMCData) | |
| p68 = TENxPBMCData("pbmc68k") | |
| rownames(p68) = make.names(rowData(p68)$Symbol_TENx, unique=TRUE) | |
| library(scater) | |
| library(scran) | |
| p68 = logNormCounts(p68) | |
| library(celldex) | |
| hpca = HumanPrimaryCellAtlasData() | |
| library(SingleR) | |
| library(BiocParallel) |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| nouns = c("cat", "mat", "cow", "chair", "table", "person", "food") | |
| pastverbs = c("sat", "walked", "rode", "spent", "moved", "wrote", "managed", "went") | |
| preps = c("on", "over", "beyond", "above", "beneath") | |
| indic = c("the", "a", "any", "some") | |
| adjectives = c("cute", "big", "small", "tiny", "huge", "green") | |
| get1 = function (x) sample(x, size=1) | |
| sentence = function() { | |
| sprintf("%s %s %s %s %s %s %s\n", get1(indic), get1(nouns), get1(pastverbs), |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| do_SingleRi = function(sce=NULL, path="/home/vincent/tenx3k.h5ad", | |
| ref=celldex::HumanPrimaryCellAtlasData(), | |
| ref.type = "label.main", | |
| min.common = 1000, assay.type.test=1L, instrument=TRUE, | |
| clprocid=NULL, ...) { | |
| if (instrument == TRUE && is.null(clprocid)) stop("clprocid not set") | |
| cl_timestamp(clprocid, "init") | |
| stopifnot(ref.type %in% c("label.main", "label.fine")) | |
| if (is.null(sce)) { | |
| is_h5ad = length(grep("h5ad$", path)==1) |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| source("inst.R", echo=TRUE) | |
| library(Rcollectl) | |
| library(AnVILBestPractices) | |
| library(SingleR) | |
| library(BiocParallel) | |
| clid = cl_start() | |
| Rcollectl::cl_timestamp(clid, "pre-data") | |
| p3k = TENxPBMCData::TENxPBMCData("pbmc3k") | |
| Rcollectl::cl_timestamp(clid, "3k loaded") | |
| rownames(p3k) = make.names(rowData(p3k)$Symbol, unique=TRUE) |