Devel 1.4.12 (#55)

* Cleaner pool iterable

* Added pipeline scripts for fast inspection

* Bugfix: SimpleFastaParser uses the entire header as identifier!

* Version++

* Even better handling of fasta files.

Expanded tests so it doesn't happen again

* Version++

* Threading (#43)

* pathos.multiprocessing -> pathos.ThreadingPool

* typo. Wrong fasta file

* Deleted comments

* Added done message

* version++

* Going back from pathos.ThreadPool to pathos.multiprocessing

* --debug: Improved logging and loggers

* version ++ 1.4.10

* Bugfix: full transition back to pathos.mp

* Removed asserts from input files: now we can use /dev/fd, stdin, stdout, devnull...

* v 1.4.11

* Fixed memory leak when multiprocessing

- Kill each child when they end their task: maxtasksperchild=1

* Version 1.4.12

README.md

@@ -6,7 +6,7 @@ Get exons from a transcriptome and raw genomic reads using abyss-bloom and bedto
 
 ## Requirements
 ```
-abyss>=2.0.0
+abyss==2.0.1  (something is happening with 2.0.2 and abyss-bloom kmers)
 bedtools (tested on 2.0)
 python3
 biopython

exfi/__init__.py

@@ -3,4 +3,4 @@
 filters.
 """
 
-__version__ = '1.4.11'
+__version__ = '1.4.12'

exfi/build_splice_graph_dict.py

@@ -191,13 +191,13 @@ def build_splice_graph_dict(
     bed6df_dict = bed3_records_to_bed6df_dict(bed3records)
 
     # Initialize pool of workers
-    pool = mp.Pool(args["threads"])
+    pool = mp.Pool(args["threads"], maxtasksperchild=1)
 
     # Build graphs in parallel and merge results
     splice_graphs = pool.map(
         build_splice_graph,
         bed6df_dict.values(),
-        chunksize=1000
+        chunksize=1000  # Number of Transcripts to process at one
     )
     pool.close()
     pool.join()

exfi/correct.py

@@ -400,14 +400,14 @@ def correct_splice_graph_dict(splice_graph_dict: SpliceGraphDict, args: dict) ->
             filled_edges_by_transcript[transcript] = set()
 
     # Initialize pool of workers
-    pool = mp.Pool(args["threads"])
+    pool = mp.Pool(args["threads"], maxtasksperchild=1)
 
     # Process each graph in parallel
     logging.info("\tCorrecting each splice graph")
     corrected_splice_graphs = pool.starmap(
         _sculpt_graph,
         zip(splice_graph_dict.values(), filled_edges_by_transcript.values()),
-        chunksize=1000
+        chunksize=1000  # Number of splice graphs to process at once.
     )
     pool.close()
     pool.join()

exfi/polish.py

@@ -153,7 +153,7 @@ def polish_splice_graph_dict(
     """
 
     # Initialize pool of workers
-    pool = mp.Pool(args["threads"])
+    pool = mp.Pool(args["threads"], maxtasksperchild=1)
 
     splice_graphs = (splice_graph for splice_graph in splice_graph_dict.values())
     fasta_dicts = (
@@ -163,7 +163,7 @@ def polish_splice_graph_dict(
     results = pool.starmap(
         polish_splice_graph,
         zip(splice_graphs, fasta_dicts),
-        chunksize=1000
+        chunksize=1000  # Number of splice_graphs to process at once
     )
     pool.close()
     pool.join()