Refactor code in notebooks

metagentools · Aug 10, 2022 · 40a9c34 · 40a9c34
1 parent ccb1d7a
commit 40a9c34
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 20 deletions.
diff --git a/notebooks/MD1_UniProt12.ipynb b/notebooks/MD1_UniProt12.ipynb
@@ -15,8 +15,19 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "from Bio import SeqIO\n",
+    "import re\n",
+    "import xlsxwriter\n",
+    "import collections\n",
     "\n",
+    "from Bio import SeqIO"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
     "contig_lengths = {}\n",
     "\n",
     "for index, record in enumerate(SeqIO.parse(project_path+\"all_genes.fna\", \"fasta\")):\n",
@@ -37,8 +48,6 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "import re\n",
-    "\n",
     "species_len = {}\n",
     "\n",
     "unclassified = []\n",
@@ -94,10 +103,6 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "import re\n",
-    "\n",
-    "min_len = 0\n",
-    "\n",
     "sorted_species = {k: v for k, v in sorted(species_len.items(), reverse=True, key=lambda item: item[1])}\n",
     "\n",
     "species_list = []\n",
@@ -112,7 +117,7 @@
     "for key, value in sorted_species.items():\n",
     "    key_strings = key.split(\" \")\n",
     "    \n",
-    "    if value > min_len and len(key_strings) > 3 and \"cellular organisms\" not in key:\n",
+    "    if len(key_strings) > 3 and \"cellular organisms\" not in key:\n",
     "        \n",
     "        taxid = key_strings[-1][:-1]\n",
     "        \n",
@@ -193,9 +198,6 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "import xlsxwriter\n",
-    "import collections\n",
-    "\n",
     "od = collections.OrderedDict(sorted(gene_species.items()))\n",
     "\n",
     "# Create a workbook and add a worksheet.\n",

diff --git a/notebooks/MD2_UniProt-12.ipynb b/notebooks/MD2_UniProt-12.ipynb
@@ -19,18 +19,10 @@
    "outputs": [],
    "source": [
     "import re\n",
-    "import csv\n",
-    "import subprocess\n",
-    "import os\n",
-    "import gzip\n",
-    "import shutil\n",
-    "import glob\n",
     "import xlsxwriter\n",
     "import collections\n",
     "\n",
-    "from datetime import datetime, timedelta\n",
-    "from Bio import SeqIO\n",
-    "from collections import defaultdict"
+    "from Bio import SeqIO"
    ]
   },
   {