From d5d6959f0a48f4ecf93ec1a9908812d8283cc982 Mon Sep 17 00:00:00 2001
From: Blaise Li <blaise.li__git@nsup.org>
Date: Fri, 1 Feb 2019 15:57:42 +0100
Subject: [PATCH] Better handle errors due to missing data.
---
small_RNA-seq/small_RNA-seq.snakefile | 120 ++++++++++++++++----------
1 file changed, 74 insertions(+), 46 deletions(-)
diff --git a/small_RNA-seq/small_RNA-seq.snakefile b/small_RNA-seq/small_RNA-seq.snakefile
index 1c38395..e532bf4 100644
--- a/small_RNA-seq/small_RNA-seq.snakefile
+++ b/small_RNA-seq/small_RNA-seq.snakefile
@@ -2400,39 +2400,51 @@ rule make_fold_heatmap:
all_folds = source_gathered_folds,
output:
fold_heatmap = OPJ(output_dir, "figures", "fold_heatmaps", "{small_type}_{fold_type}_heatmap.pdf"),
+ log:
+ warnings = OPJ(log_dir, "make_fold_heatmap", "{small_type}_{fold_type}.warnings"),
benchmark:
OPJ(log_dir, "make_fold_heatmap", "{small_type}_{fold_type}_benchmark.txt"),
threads: 16 # to limit memory usage, actually
resources:
mem_mb=8150
run:
- usetex = mpl.rcParams.get("text.usetex", False)
- all_folds = pd.read_table(input.all_folds, index_col=["gene"])
- #gene_colours = all_folds.small_type.map(small_type2colour)
- # Faster
- gene_colours = all_folds.apply(small_type_colour_setter, axis=1)
- if usetex:
- gene_colours.name = texscape(f"{wildcards.small_type}_{wildcards.fold_type}")
- else:
- gene_colours.name = f"{wildcards.small_type}_{wildcards.fold_type}"
- #all_folds.drop(["cosmid", "name", "small_type"], axis=1)
- #all_folds[all_folds.columns.difference(["cosmid", "name", "small_type"])]
- # https://github.com/mwaskom/seaborn/issues/1262
- #mpl.use("agg")
- try:
- #labels_dict = merge_with(tuple, small_type2colour, DefaultCounter(all_folds.small_type))
- # https://stackoverflow.com/a/47396625/1878788
- small_type_counts = Counter(small_type2colour.keys())
- small_type_counts.update(all_folds.small_type)
- labels_dict = {small_type : (small_type2colour[small_type], count - 1) for (small_type, count) in small_type_counts.items()}
- save_plot(
- output.fold_heatmap, plot_fold_heatmap,
- all_folds.drop(["cosmid", "name", "small_type"], axis=1),
- gene_colours, labels_dict, tight=False, rasterize=True)
- except ValueError as e:
- print(labels_dict)
- raise
- #mpl.use("PDF")
+ with warn_context(log.warnings) as warn:
+ usetex = mpl.rcParams.get("text.usetex", False)
+ all_folds = pd.read_table(input.all_folds, index_col=["gene"])
+ #gene_colours = all_folds.small_type.map(small_type2colour)
+ # Faster
+ gene_colours = all_folds.apply(small_type_colour_setter, axis=1)
+ if usetex:
+ gene_colours.name = texscape(f"{wildcards.small_type}_{wildcards.fold_type}")
+ else:
+ gene_colours.name = f"{wildcards.small_type}_{wildcards.fold_type}"
+ #all_folds.drop(["cosmid", "name", "small_type"], axis=1)
+ #all_folds[all_folds.columns.difference(["cosmid", "name", "small_type"])]
+ # https://github.com/mwaskom/seaborn/issues/1262
+ #mpl.use("agg")
+ try:
+ #labels_dict = merge_with(tuple, small_type2colour, DefaultCounter(all_folds.small_type))
+ # https://stackoverflow.com/a/47396625/1878788
+ small_type_counts = Counter(small_type2colour.keys())
+ small_type_counts.update(all_folds.small_type)
+ labels_dict = {small_type : (small_type2colour[small_type], count - 1) for (small_type, count) in small_type_counts.items()}
+ save_plot(
+ output.fold_heatmap, plot_fold_heatmap,
+ all_folds.drop(["cosmid", "name", "small_type"], axis=1),
+ gene_colours, labels_dict, tight=False, rasterize=True)
+ except ValueError as err:
+ if str(err) == "The number of observations cannot be determined on an empty distance matrix.":
+ warn("\n".join([
+ "Got ValueError:",
+ f"{str(err)}",
+ f"No data to plot in {output.fold_heatmap}\n"]))
+ warn("Generating empty file.\n")
+ # Make the file empty
+ open(output.fold_heatmap, "w").close()
+ else:
+ print(labels_dict)
+ raise
+ #mpl.use("PDF")
#def plot_norm_counts(counts_data, summaries):
@@ -3822,30 +3834,46 @@ rule make_gene_list_lfc_boxplots:
output:
boxplots = OPJ(output_dir, "figures", "{contrast}",
"{contrast}_{small_type}_{fold_type}_{gene_list}_boxplots.pdf")
+ log:
+ warnings = OPJ(log_dir, "make_gene_list_lfc_boxplots", "{contrast}_{small_type}_{fold_type}_{gene_list}.warnings"),
params:
id_lists = set_id_lists,
run:
- lfcs_dict = {}
- lfc_data = pd.read_table(input.data, index_col="gene")
- for (list_name, id_list) in params.id_lists.items():
+ with warn_context(log.warnings) as warn:
+ lfcs_dict = {}
+ lfc_data = pd.read_table(input.data, index_col="gene")
+ for (list_name, id_list) in params.id_lists.items():
+ try:
+ selected_rows = lfc_data.loc[lfc_data.index.intersection(id_list)]
+ except TypeError as err:
+ print(params.id_lists)
+ print(type(id_list))
+ print(lfc_data.index.intersection(id_list))
+ raise
+ selected_data = selected_rows[wildcards.fold_type]
+ lfcs_dict[list_name] = selected_data
+ lfcs = pd.DataFrame(lfcs_dict)
+ # lfcs = pd.DataFrame(
+ # {list_name : lfc_data.loc[set(id_list)][wildcards.fold_type] for (
+ # list_name, id_list) in params.id_lists.items()})
+ title = f"{wildcards.small_type} folds for {wildcards.contrast}"
try:
- selected_rows = lfc_data.loc[lfc_data.index.intersection(id_list)]
+ save_plot(
+ output.boxplots,
+ plot_boxplots,
+ lfcs, wildcards.fold_type,
+ title=title)
except TypeError as err:
- print(params.id_lists)
- print(type(id_list))
- print(lfc_data.index.intersection(id_list))
- raise
- selected_data = selected_rows[wildcards.fold_type]
- lfcs_dict[list_name] = selected_data
- lfcs = pd.DataFrame(lfcs_dict)
- # lfcs = pd.DataFrame(
- # {list_name : lfc_data.loc[set(id_list)][wildcards.fold_type] for (
- # list_name, id_list) in params.id_lists.items()})
- save_plot(
- output.boxplots,
- plot_boxplots,
- lfcs, wildcards.fold_type,
- title=f"{wildcards.small_type} folds for {wildcards.contrast}")
+ if str(err) == "Empty 'DataFrame': no numeric data to plot":
+ warn("\n".join([
+ "Got TypeError:",
+ f"{str(err)}",
+ f"No data to plot for {title}\n"]))
+ warn("Generating empty file.\n")
+ # Make the file empty
+ open(output.boxplots, "w").close()
+ else:
+ raise
rule make_contrast_lfc_boxplots:
--
GitLab