diff --git a/.renku/workflow/66f5307ec9194324b160ab1fd466e9d6.cwl b/.renku/workflow/66f5307ec9194324b160ab1fd466e9d6.cwl new file mode 100644 index 0000000000000000000000000000000000000000..8a0c1ca48375e3864462afa213879353525adfc9 --- /dev/null +++ b/.renku/workflow/66f5307ec9194324b160ab1fd466e9d6.cwl @@ -0,0 +1,120 @@ +class: Workflow +cwlVersion: v1.0 +hints: [] +inputs: + input_1: + default: out_folder + streamable: false + type: string + input_2: + default: data/covidtracking + streamable: false + type: string + input_3: + default: + class: File + path: ../../notebooks/process/download-covidtracking-data.ipynb + streamable: false + type: File + input_4: + default: runs/download-covidtracking-data.runs.ipynb + streamable: false + type: string + input_5: + default: states-metadata.json + streamable: false + type: string + input_6: + default: states-daily.json + streamable: false + type: string +outputs: + output_2: + outputSource: step_1/output_1 + streamable: false + type: Directory + output_3: + outputSource: step_1/output_0 + streamable: false + type: File +requirements: [] +steps: + step_1: + in: + input_1: input_1 + input_2: input_2 + input_3: input_3 + input_4: input_4 + out: + - output_0 + - output_1 + run: a17d560c41a54f5aa307ce5f3c5effe5_papermill.cwl + step_2: + in: + filename: input_5 + input_directory: step_1/output_1 + out: + - output_file + run: + arguments: [] + baseCommand: + - 'true' + class: CommandLineTool + cwlVersion: v1.0 + hints: [] + inputs: + filename: + default: states-metadata.json + streamable: false + type: string + input_directory: + streamable: false + type: Directory + outputs: + output_file: + outputBinding: + glob: $(inputs.filename) + streamable: false + type: File + permanentFailCodes: [] + requirements: + - &id001 + class: InlineJavascriptRequirement + - &id002 + class: InitialWorkDirRequirement + listing: $(inputs.input_directory.listing) + successCodes: [] + temporaryFailCodes: [] + step_3: + in: + filename: input_6 + input_directory: step_1/output_1 + out: + - output_file + run: + arguments: [] + baseCommand: + - 'true' + class: CommandLineTool + cwlVersion: v1.0 + hints: [] + inputs: + filename: + default: states-daily.json + streamable: false + type: string + input_directory: + streamable: false + type: Directory + outputs: + output_file: + outputBinding: + glob: $(inputs.filename) + streamable: false + type: File + permanentFailCodes: [] + requirements: + - *id001 + - *id002 + successCodes: [] + temporaryFailCodes: [] diff --git a/data/covidtracking/states-daily.json b/data/covidtracking/states-daily.json index 8b8ad50a5fc04493153490e89c18c634d25aa820..1f039ac48b4730222085fae250ff4cc269f88438 100644 --- a/data/covidtracking/states-daily.json +++ b/data/covidtracking/states-daily.json @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:20d05d8bf54fbe25d465293949d98baa8217af8455274ebf331eaa9fdecfc7b5 -size 3896218 +oid sha256:8cd217cca3afcedb310e36a7ea42508c92271044acc349992bea5abf79a05f6a +size 3940450 diff --git a/data/covidtracking/states-metadata.json b/data/covidtracking/states-metadata.json index 82c95999ba447b51623e9410a48a12ff64ab2dd5..28ce6ebec8dbbdb3c80f0481e0268b8eb246533a 100644 --- a/data/covidtracking/states-metadata.json +++ b/data/covidtracking/states-metadata.json @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:106cf31b2776e3d9c810972b8819576e6b9069017c01c9baf326178128121813 -size 34737 +oid sha256:8b0c5cfa8aedded7e57c76535faaa4f2b863b26d50909e984a39bbf6e8a5de59 +size 34797 diff --git a/runs/download-covidtracking-data.runs.ipynb b/runs/download-covidtracking-data.runs.ipynb index 7b7ab2ac3883501b755bfbe0f825062a7f5b3774..07cba5714688c37e81c12101d39897afdc54532e 100644 --- a/runs/download-covidtracking-data.runs.ipynb +++ b/runs/download-covidtracking-data.runs.ipynb @@ -5,10 +5,10 @@ "execution_count": 1, "metadata": { "papermill": { - "duration": 3.35233, - "end_time": "2020-06-03T15:45:09.318776", + "duration": 4.627897, + "end_time": "2020-06-04T08:28:04.985732", "exception": false, - "start_time": "2020-06-03T15:45:05.966446", + "start_time": "2020-06-04T08:28:00.357835", "status": "completed" }, "tags": [] @@ -25,10 +25,10 @@ "execution_count": 2, "metadata": { "papermill": { - "duration": 0.0259, - "end_time": "2020-06-03T15:45:09.361558", + "duration": 0.022291, + "end_time": "2020-06-04T08:28:05.020399", "exception": false, - "start_time": "2020-06-03T15:45:09.335658", + "start_time": "2020-06-04T08:28:04.998108", "status": "completed" }, "tags": [ @@ -46,10 +46,10 @@ "execution_count": 3, "metadata": { "papermill": { - "duration": 0.023503, - "end_time": "2020-06-03T15:45:09.397029", + "duration": 0.022545, + "end_time": "2020-06-04T08:28:05.052282", "exception": false, - "start_time": "2020-06-03T15:45:09.373526", + "start_time": "2020-06-04T08:28:05.029737", "status": "completed" }, "tags": [ @@ -59,7 +59,7 @@ "outputs": [], "source": [ "# Parameters\n", - "PAPERMILL_INPUT_PATH = \"/tmp/eqrts4dp/notebooks/process/download-covidtracking-data.ipynb\"\n", + "PAPERMILL_INPUT_PATH = \"/tmp/m_83rm_4/notebooks/process/download-covidtracking-data.ipynb\"\n", "PAPERMILL_OUTPUT_PATH = \"runs/download-covidtracking-data.runs.ipynb\"\n", "out_folder = \"data/covidtracking\"\n" ] @@ -68,10 +68,10 @@ "cell_type": "markdown", "metadata": { "papermill": { - "duration": 0.006779, - "end_time": "2020-06-03T15:45:09.411335", + "duration": 0.007006, + "end_time": "2020-06-04T08:28:05.068409", "exception": false, - "start_time": "2020-06-03T15:45:09.404556", + "start_time": "2020-06-04T08:28:05.061403", "status": "completed" }, "tags": [] @@ -87,10 +87,10 @@ "execution_count": 4, "metadata": { "papermill": { - "duration": 0.597563, - "end_time": "2020-06-03T15:45:10.014708", + "duration": 0.381514, + "end_time": "2020-06-04T08:28:05.457090", "exception": false, - "start_time": "2020-06-03T15:45:09.417145", + "start_time": "2020-06-04T08:28:05.075576", "status": "completed" }, "tags": [] @@ -107,10 +107,10 @@ "execution_count": 5, "metadata": { "papermill": { - "duration": 0.03125, - "end_time": "2020-06-03T15:45:10.062585", + "duration": 0.027088, + "end_time": "2020-06-04T08:28:05.496958", "exception": false, - "start_time": "2020-06-03T15:45:10.031335", + "start_time": "2020-06-04T08:28:05.469870", "status": "completed" }, "tags": [] @@ -129,10 +129,10 @@ "execution_count": 6, "metadata": { "papermill": { - "duration": 0.091143, - "end_time": "2020-06-03T15:45:10.164517", + "duration": 0.072587, + "end_time": "2020-06-04T08:28:05.580434", "exception": false, - "start_time": "2020-06-03T15:45:10.073374", + "start_time": "2020-06-04T08:28:05.507847", "status": "completed" }, "tags": [] @@ -242,10 +242,10 @@ "cell_type": "markdown", "metadata": { "papermill": { - "duration": 0.008537, - "end_time": "2020-06-03T15:45:10.188145", + "duration": 0.008248, + "end_time": "2020-06-04T08:28:05.598828", "exception": false, - "start_time": "2020-06-03T15:45:10.179608", + "start_time": "2020-06-04T08:28:05.590580", "status": "completed" }, "tags": [] @@ -259,10 +259,10 @@ "execution_count": 7, "metadata": { "papermill": { - "duration": 6.860154, - "end_time": "2020-06-03T15:45:17.054999", + "duration": 0.154111, + "end_time": "2020-06-04T08:28:05.761833", "exception": false, - "start_time": "2020-06-03T15:45:10.194845", + "start_time": "2020-06-04T08:28:05.607722", "status": "completed" }, "tags": [] @@ -279,10 +279,10 @@ "execution_count": 8, "metadata": { "papermill": { - "duration": 0.036298, - "end_time": "2020-06-03T15:45:17.109685", + "duration": 0.029699, + "end_time": "2020-06-04T08:28:05.803127", "exception": false, - "start_time": "2020-06-03T15:45:17.073387", + "start_time": "2020-06-04T08:28:05.773428", "status": "completed" }, "tags": [] @@ -301,10 +301,10 @@ "execution_count": 9, "metadata": { "papermill": { - "duration": 0.224468, - "end_time": "2020-06-03T15:45:17.346605", + "duration": 0.20856, + "end_time": "2020-06-04T08:28:06.022260", "exception": false, - "start_time": "2020-06-03T15:45:17.122137", + "start_time": "2020-06-04T08:28:05.813700", "status": "completed" }, "tags": [] @@ -314,7 +314,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "5001 data points\n" + "5057 data points\n" ] }, { @@ -364,21 +364,21 @@ " <tbody>\n", " <tr>\n", " <th>0</th>\n", - " <td>20200602</td>\n", + " <td>20200603</td>\n", " <td>AK</td>\n", - " <td>487.0</td>\n", - " <td>55716.0</td>\n", + " <td>505.0</td>\n", + " <td>57677.0</td>\n", " <td>NaN</td>\n", - " <td>10.0</td>\n", + " <td>11.0</td>\n", " <td>NaN</td>\n", " <td>NaN</td>\n", " <td>NaN</td>\n", " <td>0.0</td>\n", " <td>...</td>\n", - " <td>56203</td>\n", + " <td>58182</td>\n", " <td>0</td>\n", " <td>0</td>\n", - " <td>8bfc576f3ce0c5438919ab2a672255d0630714ec</td>\n", + " <td>02548c93014e510ca02e0fc92122c299e4aa8e78</td>\n", " <td>0</td>\n", " <td>0</td>\n", " <td>0</td>\n", @@ -388,21 +388,21 @@ " </tr>\n", " <tr>\n", " <th>1</th>\n", - " <td>20200602</td>\n", + " <td>20200603</td>\n", " <td>AL</td>\n", - " <td>18642.0</td>\n", - " <td>205160.0</td>\n", + " <td>18851.0</td>\n", + " <td>212743.0</td>\n", " <td>NaN</td>\n", " <td>NaN</td>\n", - " <td>1879.0</td>\n", + " <td>1900.0</td>\n", " <td>NaN</td>\n", " <td>591.0</td>\n", " <td>NaN</td>\n", " <td>...</td>\n", - " <td>223802</td>\n", - " <td>5</td>\n", - " <td>23</td>\n", - " <td>e925d317dd6bba774728abaf5db247f3c5983394</td>\n", + " <td>231594</td>\n", + " <td>2</td>\n", + " <td>21</td>\n", + " <td>fe5df10e9c8639e15d2ce02d16d678e16f6f6cc7</td>\n", " <td>0</td>\n", " <td>0</td>\n", " <td>0</td>\n", @@ -417,20 +417,20 @@ ], "text/plain": [ " date state positive negative pending hospitalizedCurrently \\\n", - "0 20200602 AK 487.0 55716.0 NaN 10.0 \n", - "1 20200602 AL 18642.0 205160.0 NaN NaN \n", + "0 20200603 AK 505.0 57677.0 NaN 11.0 \n", + "1 20200603 AL 18851.0 212743.0 NaN NaN \n", "\n", " hospitalizedCumulative inIcuCurrently inIcuCumulative \\\n", "0 NaN NaN NaN \n", - "1 1879.0 NaN 591.0 \n", + "1 1900.0 NaN 591.0 \n", "\n", " onVentilatorCurrently ... posNeg deathIncrease hospitalizedIncrease \\\n", - "0 0.0 ... 56203 0 0 \n", - "1 NaN ... 223802 5 23 \n", + "0 0.0 ... 58182 0 0 \n", + "1 NaN ... 231594 2 21 \n", "\n", " hash commercialScore \\\n", - "0 8bfc576f3ce0c5438919ab2a672255d0630714ec 0 \n", - "1 e925d317dd6bba774728abaf5db247f3c5983394 0 \n", + "0 02548c93014e510ca02e0fc92122c299e4aa8e78 0 \n", + "1 fe5df10e9c8639e15d2ce02d16d678e16f6f6cc7 0 \n", "\n", " negativeRegularScore negativeScore positiveScore score grade \n", "0 0 0 0 0 \n", @@ -470,18 +470,18 @@ "version": "3.7.3" }, "papermill": { - "duration": 13.234479, - "end_time": "2020-06-03T15:45:17.775676", + "duration": 7.919109, + "end_time": "2020-06-04T08:28:06.447225", "environment_variables": {}, "exception": null, - "input_path": "/tmp/eqrts4dp/notebooks/process/download-covidtracking-data.ipynb", + "input_path": "/tmp/m_83rm_4/notebooks/process/download-covidtracking-data.ipynb", "output_path": "runs/download-covidtracking-data.runs.ipynb", "parameters": { - "PAPERMILL_INPUT_PATH": "/tmp/eqrts4dp/notebooks/process/download-covidtracking-data.ipynb", + "PAPERMILL_INPUT_PATH": "/tmp/m_83rm_4/notebooks/process/download-covidtracking-data.ipynb", "PAPERMILL_OUTPUT_PATH": "runs/download-covidtracking-data.runs.ipynb", "out_folder": "data/covidtracking" }, - "start_time": "2020-06-03T15:45:04.541197", + "start_time": "2020-06-04T08:27:58.528116", "version": "1.1.0" } },