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

adding createBatchCodeTableByCountry()

Browse Source
This commit is contained in:
frankknoll
2022-02-08 13:06:16 +01:00
parent be4641d3ff
commit 1f1bdd9293
1 changed files with 103 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

105
HowBadIsMyBatch.ipynb
View File

@@ -149,6 +149,9 @@
" def filterByCovid19(self, dataFrame):\n",
" return dataFrame[self._isCovid19(dataFrame)]\n",
"\n",
" def filterByCountry(self, dataFrame, country, countryColumnName):\n",
" return dataFrame[dataFrame[countryColumnName] == country]\n",
"\n",
" def filterBy(self, dataFrame, manufacturer = None, dose = None):\n",
" return dataFrame[self._isManufacturer(dataFrame, manufacturer) & self._isDose(dataFrame, dose)]\n",
"\n",
@@ -278,6 +281,7 @@
" dataFrame['VAX_DOSE_SERIES'].rename('Dose')\n",
" ]))\n",
"\n",
" # FK-TODO: rename _createDoseTable()\n",
" @staticmethod\n",
" def _createDoseTable(dataFrame):\n",
" doseTable = SummationTableFactory.createSummationTable(\n",
@@ -311,12 +315,25 @@
" return internationalLotTable.sort_values(by = 'Severe reports (%)', ascending = False)\n",
"\n",
" @staticmethod\n",
" def createBatchCodeTableByCountry(dataFrame : pd.DataFrame, country):\n",
" dataFrame = DataFrameFilter().filterByCovid19(dataFrame)\n",
" batchCodeTable = InternationalLotTableFactory._createBatchCodeTableByCountry(dataFrame, country)\n",
" return batchCodeTable.sort_values(by = 'Severe reports (%)', ascending = False)\n",
"\n",
" @staticmethod\n",
" def _createInternationalLotTable(dataFrame):\n",
" countryColumnName = 'Country'\n",
" InternationalLotTableFactory._addCountryColumn(dataFrame, countryColumnName = countryColumnName)\n",
" dataFrame = InternationalLotTableFactory._addCountryColumn(dataFrame, countryColumnName = countryColumnName)\n",
" return DoseTableFactory._createDoseTable(dataFrame.groupby(dataFrame[countryColumnName]))\n",
"\n",
" @staticmethod\n",
" def _createBatchCodeTableByCountry(dataFrame : pd.DataFrame, country):\n",
" countryColumnName = 'Country'\n",
" dataFrame = InternationalLotTableFactory._addCountryColumn(dataFrame, countryColumnName = countryColumnName)\n",
" dataFrame = DataFrameFilter().filterByCountry(dataFrame, country = country, countryColumnName = countryColumnName)\n",
" return DoseTableFactory._createDoseTable(dataFrame.groupby('VAX_LOT'))\n",
"\n",
" @staticmethod\n",
" def _addCountryColumn(dataFrame, countryColumnName):\n",
" dataFrame[countryColumnName] = dataFrame.apply(\n",
" lambda row:\n",
@@ -324,6 +341,7 @@
" splttype = row['SPLTTYPE'],\n",
" default = 'Unknown Country'),\n",
" axis = 'columns')\n",
" return dataFrame.astype({countryColumnName: \"string\"})\n",
"\n",
" @staticmethod\n",
" def _getCountryNameOfSplttypeOrDefault(splttype, default):\n",
@@ -810,7 +828,40 @@
" 'United Kingdom',\n",
" 'Unknown Country'\n",
" ],\n",
" name = 'Country')))\n"
" dtype = \"string\",\n",
" name = 'Country')))\n",
"\n",
" def test_createBatchCodeTableByCountry(self):\n",
" # Given\n",
" dataFrame = TestHelper.createDataFrame(\n",
" columns = ['DIED', 'L_THREAT', 'DISABLE', 'VAX_TYPE', 'VAX_MANU', 'VAX_LOT', 'VAX_DOSE_SERIES', 'SPLTTYPE', 'HOSPITAL', 'ER_VISIT'],\n",
" data = [ [1, 0, 0, 'COVID19', 'MODERNA', '016M20A', '2', 'GBPFIZER INC2020486806', 0, 0],\n",
" [1, 0, 0, 'COVID19', 'MODERNA', '030L20A', '1', 'FRMODERNATX, INC.MOD20224', 0, 0],\n",
" [1, 1, 1, 'COVID19', 'MODERNA', '030L20B', '1', 'FRMODERNATX, INC.MOD20224', 0, 0],\n",
" [0, 1, 1, 'COVID19', 'MODERNA', '030L20B', '1', 'FRMODERNATX, INC.MOD20224', 0, 0]],\n",
" index = [\n",
" \"1048786\",\n",
" \"1048786\",\n",
" \"4711\",\n",
" \"0815\"])\n",
" \n",
" # When\n",
" batchCodeTable = InternationalLotTableFactory.createBatchCodeTableByCountry(dataFrame, 'France')\n",
"\n",
" # Then\n",
" assert_frame_equal(\n",
" batchCodeTable,\n",
" TestHelper.createDataFrame(\n",
" columns = ['Total reports', 'Deaths', 'Disabilities', 'Life Threatening Illnesses', 'Severe reports (%)'],\n",
" data = [ [2, 1, 2, 2, (1 + 2 + 2) / 2 * 100],\n",
" [1, 1, 0, 0, (1 + 0 + 0) / 1 * 100]],\n",
" index = pd.Index(\n",
" [\n",
" '030L20B',\n",
" '030L20A'\n",
" ],\n",
" name = 'VAX_LOT')),\n",
" check_dtype = False)\n"
]
},
{
@@ -1028,6 +1079,56 @@
"internationalLotTable.to_excel('results/International_Deadly_Lots.xlsx')\n",
"internationalLotTable"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "0f691bf5",
"metadata": {},
"outputs": [],
"source": [
"def createBatchCodeTableByCountry(country):\n",
" vaersDescr = VaersDescrReader(dataDir = 'VAERS').readNonDomesticVaersDescr()\n",
" dataFrame = VaersDescr2DataFrameConverter.createDataFrameFromDescr(vaersDescr)\n",
" DataFrameNormalizer.normalize(dataFrame)\n",
" return InternationalLotTableFactory.createBatchCodeTableByCountry(dataFrame, country)\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "59f7e62c",
"metadata": {},
"outputs": [],
"source": [
"batchCodeTable = createBatchCodeTableByCountry('France')"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "a35ede95",
"metadata": {},
"outputs": [],
"source": [
"import os\n",
"\n",
"report_path = 'results/international/'\n",
"if not os.path.exists(report_path):\n",
" os.makedirs(report_path)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "ff259a35",
"metadata": {},
"outputs": [],
"source": [
"batchCodeTable = batchCodeTable[batchCodeTable['Total reports'] > 50]\n",
"batchCodeTable.to_excel(report_path + 'france.xlsx')\n",
"batchCodeTable"
]
}
],
"metadata": {