More benchmarking

additional_benchmarks/Snakefile

@@ -71,11 +71,35 @@ rule plyranges_join_overlap_inner:
  {params.rscript} Rscripts/plyranges_join_overlap_inner.r {input.a} {input.b} > {output}
  """
 
-#tools = ["bedtools_intersect", "granges_filter", "plyranges_join_overlap_inner"]
 tools = ["bedtools_intersect", "granges_filter"]
 sizes = np.logspace(3, 7, 10).astype('int')
 
-all_benches = expand("results/{tool}__size_{size}.bed", tool=tools, size=sizes)
+all_results = expand("results/{tool}__size_{size}.bed", tool=tools, size=sizes)
 
 rule all:
- input: all_benches
+ input: all_results
+
+
+rule combine:
+ input: expand("benchmarks/{tool}__size_{size}.tsv", tool=tools, size=sizes)
+ output: "combined_benchmarks.tsv"
+ run:
+ import pandas as pd
+
+ def read_tsv(file):
+ return pd.read_csv(file, sep="\t")
+
+ dfs = []
+ for tool in tools:
+ for size in sizes:
+ file = f"benchmarks/{tool}__size_{size}.tsv"
+ df = read_tsv(file)
+ df["tool"] = tool
+ df["size"] = size
+ dfs.append(df)
+
+ combined_df = pd.concat(dfs, ignore_index=True)
+ combined_df.to_csv(output[0], sep="\t", index=False)
+
+rule merge:
+ input: "combined_benchmarks.tsv"