Fix linting errors, minor bugs and reorganise code

* Fix some linting errors and some minor bugs caught by the linter.
  Move the `random_string` function to separate module for use in multiple
  places in the code.

5 files changed, 52 insertions, 23 deletions

diff --git a/gn3/computations/heatmap.py b/gn3/computations/heatmap.py
index 92014cf..1143450 100644
--- a/gn3/computations/heatmap.py
+++ b/gn3/computations/heatmap.py
@@ -6,6 +6,7 @@ generate various kinds of heatmaps.
 from functools import reduce
 from typing import Any, Dict, Sequence
 from gn3.computations.slink import slink
+from gn3.computations.qtlreaper import generate_traits_file
 from gn3.computations.correlations2 import compute_correlation
 from gn3.db.genotypes import build_genotype_file, load_genotype_samples
 from gn3.db.traits import (
@@ -155,14 +156,14 @@ def heatmap_data(traits_names, conn: Any):
         for fullname in traits_names]
     traits_list = tuple(x[0] for x in traits_details)
     traits_data_list = [x[1] for x in traits_details]
-    exported_traits_data_list = tuple(
-        export_trait_data(td, strainlist) for td in traits_data_list)
     genotype_filename = build_genotype_file(traits_list[0]["riset"])
     strainlist = load_genotype_samples(genotype_filename)
+    exported_traits_data_list = tuple(
+        export_trait_data(td, strainlist) for td in traits_data_list)
     slink_data = slink(cluster_traits(exported_traits_data_list))
     ordering_data = compute_heatmap_order(slink_data)
     strains_and_values = retrieve_strains_and_values(
-        orders, strainlist, exported_traits_data_list)
+        ordering_data, strainlist, exported_traits_data_list)
     strains_values = strains_and_values[0][1]
     trait_values = [t[2] for t in strains_and_values]
     traits_filename = generate_traits_filename()
diff --git a/gn3/computations/qtlreaper.py b/gn3/computations/qtlreaper.py
index 3b8e4db..30c7051 100644
--- a/gn3/computations/qtlreaper.py
+++ b/gn3/computations/qtlreaper.py
@@ -3,17 +3,10 @@ This module contains functions to interact with the `qtlreaper` utility for
 computation of QTLs.
 """
 import os
-import random
-import string
 import subprocess
+from gn3.random import random_string
 from gn3.settings import TMPDIR, REAPER_COMMAND
 
-def random_string(length):
-    """Generate a random string of length `length`."""
-    return "".join(
-        random.choices(
-            string.ascii_letters + string.digits, k=length))
-
 def generate_traits_file(strains, trait_values, traits_filename):
     """
     Generate a traits file for use with `qtlreaper`.
@@ -25,11 +18,13 @@ def generate_traits_file(strains, trait_values, traits_filename):
         computation of QTLs.
     """
     header = "Trait\t{}\n".format("\t".join(strains))
-    data = [header] + [
-        "T{}\t{}\n".format(i+1, "\t".join([str(i) for i in t]))
-        for i, t in enumerate(trait_values[:-1])] + [
-        "T{}\t{}".format(len(trait_values), "\t".join([str(i) for i in t]))
-        for t in trait_values[-1:]]
+    data = (
+        [header] +
+        ["T{}\t{}\n".format(i+1, "\t".join([str(i) for i in t]))
+         for i, t in enumerate(trait_values[:-1])] +
+        ["T{}\t{}".format(
+            len(trait_values), "\t".join([str(i) for i in t]))
+         for t in trait_values[-1:]])
     with open(traits_filename, "w") as outfile:
         outfile.writelines(data)
 
@@ -93,6 +88,9 @@ def run_reaper(
 
 
 def parse_reaper_main_results(results_file):
+    """
+    Parse the results file of running QTLReaper into a list of dicts.
+    """
     with open(results_file, "r") as infile:
         lines = infile.readlines()
 
@@ -104,6 +102,9 @@ def parse_reaper_main_results(results_file):
     return [dict(zip(header, __parse_line(line))) for line in lines[1:]]
 
 def parse_reaper_permutation_results(results_file):
+    """
+    Parse the results QTLReaper permutations into a list of values.
+    """
     with open(results_file, "r") as infile:
         lines = infile.readlines()
 
diff --git a/gn3/db/traits.py b/gn3/db/traits.py
index ccb101a..bfe887e 100644
--- a/gn3/db/traits.py
+++ b/gn3/db/traits.py
@@ -1,6 +1,8 @@
 """This class contains functions relating to trait data manipulation"""
-from gn3.settings import TMPDIR
+import os
 from typing import Any, Dict, Union, Sequence
+from gn3.settings import TMPDIR
+from gn3.random import random_string
 from gn3.function_helpers import compose
 from gn3.db.datasets import retrieve_trait_dataset
 
@@ -669,5 +671,6 @@ def retrieve_trait_data(trait: dict, conn: Any, strainlist: Sequence[str] = tupl
     return {}
 
 def generate_traits_filename(base_path: str = TMPDIR):
+    """Generate a unique filename for use with generated traits files."""
     return "{}/traits_test_file_{}.txt".format(
         os.path.abspath(base_path), random_string(10))
diff --git a/gn3/heatmaps/heatmaps.py b/gn3/heatmaps/heatmaps.py
index 3bf7917..88f546d 100644
--- a/gn3/heatmaps/heatmaps.py
+++ b/gn3/heatmaps/heatmaps.py
@@ -14,6 +14,19 @@ def generate_random_data(data_stop: float = 2, width: int = 10, height: int = 30
     return [[random.uniform(0,data_stop) for i in range(0, width)]
             for j in range(0, height)]
 
+def generate_random_data2(data_stop: float = 2, width: int = 10, height: int = 30):
+    """
+    This is mostly a utility function to be used to generate random data, useful
+    for development of the heatmap generation code, without access to the actual
+    database data.
+    """
+    return [
+        [{
+            "value": item,
+            "category": random.choice(["C57BL/6J +", "DBA/2J +"])}
+         for item in axis]
+        for axis in generate_random_data(data_stop, width, height)]
+
 def heatmap_x_axis_names():
     return [
         "UCLA_BXDBXH_CARTILAGE_V2::ILM103710672",
@@ -30,13 +43,14 @@ def heatmap_x_axis_names():
 
 # Grey + Blue + Red
 def generate_heatmap():
-    rows = 20
-    data = generate_random_data(height=rows)
-    y = (["%s"%x for x in range(1, rows+1)][:-1] + ["X"]) #replace last item with x for now
+    cols = 20
+    y_axis = (["%s"%x for x in range(1, cols+1)][:-1] + ["X"]) #replace last item with x for now
+    x_axis = heatmap_x_axis_names()
+    data = generate_random_data(height=cols, width=len(x_axis))
     fig = px.imshow(
         data,
-        x=heatmap_x_axis_names(),
-        y=y,
+        x=x_axis,
+        y=y_axis,
         width=500)
     fig.update_traces(xtype="array")
     fig.update_traces(ytype="array")
@@ -49,6 +63,5 @@ def generate_heatmap():
         coloraxis_colorscale=[
             [0.0, '#3B3B3B'], [0.4999999999999999, '#ABABAB'],
             [0.5, '#F5DE11'], [1.0, '#FF0D00']])
-
     fig.write_html("%s/%s"%(heatmap_dir, "test_image.html"))
     return fig
diff --git a/gn3/random.py b/gn3/random.py
new file mode 100644
index 0000000..f0ba574
--- /dev/null
+++ b/gn3/random.py
@@ -0,0 +1,11 @@
+"""
+Functions to generate complex random data.
+"""
+import random
+import string
+
+def random_string(length):
+    """Generate a random string of length `length`."""
+    return "".join(
+        random.choices(
+            string.ascii_letters + string.digits, k=length))