diff options
| -rw-r--r-- | gn3/api/rqtl.py | 49 | ||||
| -rw-r--r-- | gn3/api/rqtl2.py | 7 | ||||
| -rw-r--r-- | gn3/computations/rqtl.py | 7 |
3 files changed, 49 insertions, 14 deletions
diff --git a/gn3/api/rqtl.py b/gn3/api/rqtl.py index 30173f9..7246ba8 100644 --- a/gn3/api/rqtl.py +++ b/gn3/api/rqtl.py @@ -1,5 +1,7 @@ """Endpoints for running the rqtl cmd""" import os +import uuid +import subprocess from pathlib import Path from flask import Blueprint @@ -13,6 +15,7 @@ from gn3.fs_helpers import assert_path_exists, get_tmpdir rqtl = Blueprint("rqtl", __name__) + @rqtl.route("/compute", methods=["POST"]) def compute(): """Given at least a geno_file and pheno_file, generate and @@ -21,17 +24,22 @@ run the rqtl_wrapper script and return the results as JSON """ genofile = request.form['geno_file'] phenofile = request.form['pheno_file'] - assert_path_exists(genofile) assert_path_exists(phenofile) - # Split kwargs by those with values and boolean ones that just convert to True/False + run_id = request.args.get("id") + with open(os.path.join(current_app.config.get("TMPDIR"), + f"{run_id}.txt"), "w+"): + pass + # Split kwargs by those with values and boolean ones + # that just convert to True/False kwargs = ["covarstruct", "model", "method", "nperm", "scale", "control"] boolean_kwargs = ["addcovar", "interval", "pstrata", "pairscan"] all_kwargs = kwargs + boolean_kwargs rqtl_kwargs = {"geno": genofile, "pheno": phenofile, "outdir": current_app.config.get("TMPDIR")} rqtl_bool_kwargs = [] + for kwarg in all_kwargs: if kwarg in request.form: if kwarg in kwargs: @@ -39,7 +47,7 @@ run the rqtl_wrapper script and return the results as JSON if kwarg in boolean_kwargs: rqtl_bool_kwargs.append(kwarg) - outdir = os.path.join(get_tmpdir(),"gn3") + outdir = os.path.join(get_tmpdir(), "gn3") if not os.path.isdir(outdir): os.mkdir(outdir) @@ -52,9 +60,13 @@ run the rqtl_wrapper script and return the results as JSON ) rqtl_output = {} + # get the stdout file + run_id = request.args.get("id", str(uuid.uuid4())) if not os.path.isfile(os.path.join(current_app.config.get("TMPDIR"), "gn3", rqtl_cmd.get('output_file'))): - os.system(rqtl_cmd.get('rqtl_cmd')) + stream_ouput_file = os.path.join(current_app.config.get("TMPDIR"), + f"{run_id}.txt") + run_process(rqtl_cmd.get("rqtl_cmd"), stream_ouput_file, run_id) if "pairscan" in rqtl_bool_kwargs: rqtl_output['results'] = process_rqtl_pairscan(rqtl_cmd.get('output_file'), genofile) @@ -62,7 +74,32 @@ run the rqtl_wrapper script and return the results as JSON rqtl_output['results'] = process_rqtl_mapping(rqtl_cmd.get('output_file')) if int(rqtl_kwargs['nperm']) > 0: - rqtl_output['perm_results'], rqtl_output['suggestive'], rqtl_output['significant'] = \ - process_perm_output(rqtl_cmd.get('output_file')) + rqtl_output['perm_results'], rqtl_output['suggestive'], + rqtl_output['significant'] = process_perm_output(rqtl_cmd.get('output_file')) return jsonify(rqtl_output) + + +def run_process(rscript_cmd, output_file, run_id): + """Main function to do the streaming""" + # TODO: move this function to own file + # pylint: disable=consider-using-with + process = subprocess.Popen( + rscript_cmd, shell=True, + stdout=subprocess.PIPE, + stderr=subprocess.STDOUT + ) + for line in iter(process.stdout.readline, b""): + # these allow endpoint stream to read the file since + # no read and write file same tiem + with open(output_file, "a+") as file_handler: + file_handler.write(line.decode("utf-8")) + process.stdout.close() + process.wait() + if process.returncode == 0: + return jsonify({"msg": "success", + "results": "file_here", + "run_id": run_id}) + return jsonify({"msg": "fail", + "error": "Process failed", + "run_id": run_id}) diff --git a/gn3/api/rqtl2.py b/gn3/api/rqtl2.py index 55d847e..a6ac411 100644 --- a/gn3/api/rqtl2.py +++ b/gn3/api/rqtl2.py @@ -57,6 +57,7 @@ def stream(identifier="output"): with open(output_file, encoding="utf-8") as file_handler: # read to the last position default to 0 file_handler.seek(seek_position) - return jsonify({"data": file_handler.readlines(), - "run_id": identifier, - "pointer": file_handler.tell()}) + results = {"data": file_handler.readlines(), + "run_id": identifier, + "pointer": file_handler.tell()} + return jsonify(results) diff --git a/gn3/computations/rqtl.py b/gn3/computations/rqtl.py index 9e39c3f..eaf5df5 100644 --- a/gn3/computations/rqtl.py +++ b/gn3/computations/rqtl.py @@ -67,8 +67,7 @@ def process_rqtl_mapping(file_name: str) -> List: # Later I should probably redo this using csv.read to avoid the # awkwardness with removing quotes with [1:-1] outdir = os.path.join(get_tmpdir(),"gn3") - - with open( os.path.join(outdir,file_name),"r",encoding="utf-8") as the_file: + with open(os.path.join(outdir,file_name),"r",encoding="utf-8") as the_file: for line in the_file: line_items = line.split(",") if line_items[1][1:-1] == "chr" or not line_items: @@ -316,15 +315,13 @@ def process_perm_output(file_name: str) -> Tuple[List, float, float]: with open(os.path.join(outdir, file_name), "r", encoding="utf-8") as the_file: + for i, line in enumerate(the_file): if i == 0: # Skip header line continue - _snp, _chromosome, _position, lod_score = line.split(",") perm_results.append(float(lod_score)) - suggestive = np.percentile(np.array(perm_results), 67) significant = np.percentile(np.array(perm_results), 95) - return perm_results, suggestive, significant |