Updated installation

.gitignore

@@ -0,0 +1,2 @@
+
+.DS_Store

assembly_environment.yml

@@ -31,7 +31,7 @@ dependencies:
   - biobakery_workflows=3.1=pyh0d95946_0
   - biom-format=2.1.10=py38h6c62de6_1
   - biopython=1.80=py38h0a891b7_0
-  - blas=1.1=openblas
+  - blas=1.0=openblas
   - blast=2.13.0=hf3cf87c_0
   - bleach=5.0.1=pyhd8ed1ab_0
   - blinker=1.5=pyhd8ed1ab_0
@@ -44,15 +44,16 @@ dependencies:
   - bx-python=0.9.0=py38h616c765_0
   - bzip2=1.0.8=h7f98852_4
   - c-ares=1.18.1=h7f98852_0
-  - ca-certificates=2022.12.7=ha878542_0
+  - ca-certificates=2023.11.17=hbcca054_0
   - cached-property=1.5.2=hd8ed1ab_1
   - cached_property=1.5.2=pyha770c72_1
   - cachetools=4.2.4=pyhd8ed1ab_0
   - cairo=1.16.0=h18b612c_1001
   - capnproto=0.10.2=h6239696_0
-  - certifi=2022.12.7=pyhd8ed1ab_0
+  - certifi=2023.11.17=pyhd8ed1ab_0
   - cffi=1.15.1=py38h4a40e3a_3
   - charset-normalizer=2.1.1=pyhd8ed1ab_0
+  - checkm2=1.0.1=pyh7cba7a3_0
   - click=8.1.3=py38h578d9bd_1
   - cloudpickle=2.2.0=pyhd8ed1ab_0
   - cmseq=1.0.4=pyhb7b1952_0
@@ -201,7 +202,7 @@ dependencies:
   - openblas=0.3.21=pthreads_h320a7e8_3
   - openjdk=11.0.1=h516909a_1016
   - openjpeg=2.5.0=h7d73246_1
-  - openssl=1.1.1s=h0b41bf4_1
+  - openssl=1.1.1w=hd590300_0
   - opt_einsum=3.3.0=pyhd8ed1ab_1
   - packaging=22.0=pyhd8ed1ab_0
   - pandas=1.4.0=py38h43a58ef_0
@@ -271,21 +272,47 @@ dependencies:
   - r-base=4.2.0=h1ae530e_0
   - r-boot=1.3_28=r42h6115d3f_0
   - r-class=7.3_20=r42h76d94ec_0
+  - r-cli=3.6.1=r42ha503ecb_1
   - r-cluster=2.1.3=r42h640688f_0
   - r-codetools=0.2_18=r42h6115d3f_0
+  - r-crayon=1.5.2=r42hc72bb7e_2
+  - r-data.table=1.14.8=r42h029312a_2
+  - r-docopt=0.7.1=r42hc72bb7e_3
+  - r-doparallel=1.0.17=r42hc72bb7e_2
+  - r-dplyr=1.1.4=r42ha503ecb_0
+  - r-ellipsis=0.3.2=r42h57805ef_2
+  - r-fansi=1.0.5=r42h57805ef_0
+  - r-foreach=1.5.2=r42hc72bb7e_2
   - r-foreign=0.8_82=r42h76d94ec_0
+  - r-generics=0.1.3=r42hc72bb7e_2
+  - r-glue=1.6.2=r42h57805ef_2
+  - r-iterators=1.0.14=r42hc72bb7e_2
   - r-kernsmooth=2.23_20=r42h640688f_0
   - r-lattice=0.20_45=r42h76d94ec_0
+  - r-lifecycle=1.0.4=r42hc72bb7e_0
+  - r-magrittr=2.0.3=r42h57805ef_2
   - r-mass=7.3_57=r42h76d94ec_0
   - r-matrix=1.4_1=r42h76d94ec_0
   - r-mgcv=1.8_40=r42h76d94ec_0
   - r-nlme=3.1_157=r42h640688f_0
   - r-nnet=7.3_17=r42h76d94ec_0
+  - r-pillar=1.9.0=r42hc72bb7e_1
+  - r-pkgconfig=2.0.3=r42hc72bb7e_3
+  - r-purrr=1.0.2=r42h57805ef_0
+  - r-r6=2.5.1=r42hc72bb7e_2
   - r-recommended=4.2=r42hd8ed1ab_1005
+  - r-rlang=1.1.2=r42ha503ecb_0
   - r-rpart=4.1.16=r42h76d94ec_0
   - r-spatial=7.3_15=r42h76d94ec_0
   - r-stringi=1.7.6=r42h884c59f_0
+  - r-stringr=1.5.1=r42h785f33e_0
   - r-survival=3.3_1=r42h76d94ec_0
+  - r-tibble=3.2.1=r42h57805ef_2
+  - r-tidyr=1.3.0=r42ha503ecb_1
+  - r-tidyselect=1.2.0=r42hc72bb7e_1
+  - r-utf8=1.2.4=r42h57805ef_0
+  - r-vctrs=0.6.5=r42ha503ecb_0
+  - r-withr=2.5.2=r42hc72bb7e_0
   - raxml=8.2.12=hec16e2b_4
   - re2=2021.04.01=h9c3ff4c_0
   - readline=8.1.2=h0f457ee_0

assembly_workflow.py

@@ -286,7 +286,7 @@ def list_depends(name, step, paired):
 			else:
 				return [str(name + pair_identifier + "." + input_extension), str(name + pair_identifier_2 + "." + input_extension)]
 		elif paired == "concatenated":
-			return [str(deconcatenated_dir + name.split("/")[-1] + ".done"), list_paired]
+			return [str(deconcatenated_dir + name.split("/")[-1] + ".done"), list_paired, scratch + "searched/paired_list.txt"]
 		else:
 			return [str(name + "." + input_extension)]
 	elif step == "align":
@@ -381,6 +381,11 @@ def deconcatenate(name):
 			command = "if grep -q -m 1 /2$ " + name + "." + input_extension + "; then echo " + name + "." + input_extension + " >> " + list_paired + "; fi && touch " + searched_dir + name.split("/")[-1] + "_searched.log"
 			workflow.add_task(command, depends=[name + "." + input_extension], targets = [searched_dir + name.split("/")[-1] + "_searched.log", list_paired], name="Create paired/unpaired list")
 
+	command = "mkdir -p " + scratch + "searched/ && cp " + list_paired + " " + scratch + "searched/paired_list.txt"
+	depends_list = [searched_dir + name.split("/")[-1] + "_searched.log" for name in names]
+	depends_list.append(list_paired)
+	workflow.add_task(command, depends=depends_list, targets = [scratch + "searched/paired_list.txt"], name="Create paired/unpaired list")
+
 	for name in names:
 		workflow.add_task_gridable(deconcatenate(name),
 			depends=list_depends(name=name, step="deconcatenate", paired="concatenated"),
@@ -871,8 +876,15 @@ def copy_bins(name):
 		command = "cat " + qa_unmerged_dir + name.split("/")[-1] + "/quality_report.tsv > " + qa_dir + "quality_report.tsv"
 		depends_list.append(qa_unmerged_dir + name.split("/")[-1] + "/quality_report.tsv")
 	else:
-		command = command + " && tail -n +2 " + qa_unmerged_dir + name.split("/")[-1] + "/quality_report.tsv >> " + qa_dir + "quality_report.tsv"
 		depends_list.append(qa_unmerged_dir + name.split("/")[-1] + "/quality_report.tsv")
+
+with open(qa_dir + "quality_report_merge_list.txt", 'w') as f:
+	for count, depend in enumerate(depends_list):
+		if count != 0:
+			f.write(depend + '\n')
+
+command = command + " && cat " + qa_dir + "quality_report_merge_list.txt" + " | while read line; do tail -n +2 $line >> " + qa_dir + "quality_report.tsv; done"
+
 workflow.add_task(command,
 depends = depends_list,
 targets = qa_dir + "quality_report.tsv",
@@ -916,8 +928,15 @@ def copy_bins(name):
 			command = "cat " + phylophlan_unmerged_dir + name.split("/")[-1] + "/" + "phylophlan_out.tsv > " + phylophlan_dir + "phylophlan_out.tsv"
 			depends_list.append(phylophlan_unmerged_dir + name.split("/")[-1] + "/" + "phylophlan_out.tsv")
 		else:
-			command = command + " && tail -n +5 " + phylophlan_unmerged_dir + name.split("/")[-1] + "/" + "phylophlan_out.tsv >> " + phylophlan_dir + "phylophlan_out.tsv"
 			depends_list.append(phylophlan_unmerged_dir + name.split("/")[-1] + "/" + "phylophlan_out.tsv")
+
+	command = command + " && cat " + phylophlan_dir + "phylophlan_merge_list.txt" + " | while read line; do tail -n +5 $line >> " + phylophlan_dir + "phylophlan_out.tsv; done"
+
+	with open(phylophlan_dir + "phylophlan_merge_list.txt", 'w') as f:
+		for count, depend in enumerate(depends_list):
+			if count != 0:
+				f.write(depend + '\n')
+
 	workflow.add_task(command,
 	depends = depends_list,
 	targets = phylophlan_dir + "phylophlan_out.tsv",