francois/HowBadIsMyBatch
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

refactoring

Browse Source
This commit is contained in:
frankknoll
2022-01-31 20:36:20 +01:00
parent 663a560c67
commit ce981220d1
1 changed files with 51 additions and 35 deletions
Download Patch File Download Diff File Expand all files Collapse all files

86
HowBadIsMyBatch.ipynb
View File

@@ -17,40 +17,7 @@
{ {
"cell_type": "code", "cell_type": "code",
"execution_count": null, "execution_count": null,
"id": "7b5d6df0", "id": "a271254b",
"metadata": {},
"outputs": [],
"source": [
"def createDataFrameFromDescr(vaersDescr):\n",
" return pd.merge(\n",
" vaersDescr['VAERSDATA'],\n",
" vaersDescr['VAERSVAX'],\n",
" how = 'left',\n",
" left_index = True,\n",
" right_index = True,\n",
" validate = 'one_to_many')\n",
"\n",
"def createDataFrameFromDescrs(vaersDescrs):\n",
" dataFrames = map(createDataFrameFromDescr, vaersDescrs)\n",
" return pd.concat(dataFrames)\n",
"\n",
"def filterDataFrame(df, manufacturer = None, dose = None):\n",
" isCovid19 = df[\"VAX_TYPE\"] == \"COVID19\"\n",
" isManufacturer = df[\"VAX_MANU\"] == manufacturer if manufacturer is not None else True\n",
" isDose = df[\"VAX_DOSE_SERIES\"].str.contains(dose) if dose is not None else True\n",
" return df[isCovid19 & isManufacturer & isDose]\n",
"\n",
"def createAndFilterDataFrameFromDescrs(vaersDescrs, manufacturer, dose):\n",
" return filterDataFrame(createDataFrameFromDescrs(vaersDescrs), manufacturer = manufacturer, dose = dose)\n",
"\n",
"def createDataFrameSevereEffectsFromDescrs(vaersDescrs, dose):\n",
" return filterDataFrame(createDataFrameFromDescrs(vaersDescrs), dose = dose)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "233bc590",
"metadata": {}, "metadata": {},
"outputs": [], "outputs": [],
"source": [ "source": [
@@ -78,8 +45,57 @@
" }\n", " }\n",
"\n", "\n",
"def readVaersDescrs(dataDir, years):\n", "def readVaersDescrs(dataDir, years):\n",
" return [readVaersDescr(dataDir, year) for year in years]\n", " return [readVaersDescr(dataDir, year) for year in years]\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "7b5d6df0",
"metadata": {},
"outputs": [],
"source": [
"def createDataFrameFromDescr(vaersDescr):\n",
" return pd.merge(\n",
" vaersDescr['VAERSDATA'],\n",
" vaersDescr['VAERSVAX'],\n",
" how = 'left',\n",
" left_index = True,\n",
" right_index = True,\n",
" validate = 'one_to_many')\n",
"\n", "\n",
"def createDataFrameFromDescrs(vaersDescrs):\n",
" dataFrames = map(createDataFrameFromDescr, vaersDescrs)\n",
" return pd.concat(dataFrames)\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "3ebcba86",
"metadata": {},
"outputs": [],
"source": [
"def filterDataFrame(df, manufacturer = None, dose = None):\n",
" isCovid19 = df[\"VAX_TYPE\"] == \"COVID19\"\n",
" isManufacturer = df[\"VAX_MANU\"] == manufacturer if manufacturer is not None else True\n",
" isDose = df[\"VAX_DOSE_SERIES\"].str.contains(dose) if dose is not None else True\n",
" return df[isCovid19 & isManufacturer & isDose]\n",
"\n",
"def createAndFilterDataFrameFromDescrs(vaersDescrs, manufacturer, dose):\n",
" return filterDataFrame(createDataFrameFromDescrs(vaersDescrs), manufacturer = manufacturer, dose = dose)\n",
"\n",
"def createDataFrameSevereEffectsFromDescrs(vaersDescrs, dose):\n",
" return filterDataFrame(createDataFrameFromDescrs(vaersDescrs), dose = dose)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "5d2d2f93",
"metadata": {},
"outputs": [],
"source": [
"def createAndFilterDataFrameFromFiles(dataDir, manufacturer, dose):\n", "def createAndFilterDataFrameFromFiles(dataDir, manufacturer, dose):\n",
" return createAndFilterDataFrameFromDescrs(readVaersDescrs(dataDir, [\"2021\", \"2022\"]), manufacturer, dose)\n", " return createAndFilterDataFrameFromDescrs(readVaersDescrs(dataDir, [\"2021\", \"2022\"]), manufacturer, dose)\n",
"\n", "\n",