diff --git a/.renku/workflow/0308e2c322734725b07608fbaf75dca6.cwl b/.renku/workflow/0308e2c322734725b07608fbaf75dca6.cwl new file mode 100644 index 0000000000000000000000000000000000000000..308b84ab386a03de4ca42017340ebf3f55f574ea --- /dev/null +++ b/.renku/workflow/0308e2c322734725b07608fbaf75dca6.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_0: + outputSource: step_1/output_1 + streamable: false + type: Directory + output_2: + 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_1 + - output_0 + 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 1d6060de34997b5d9b052d1d7fde5f1a11de941b..7201bef2539f5bd665eb01ba8eaa9f3eb27a0ef9 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 sha256:8559ed3a981bb365de2dbd4f0293ec03b39c5b1a4754a94d59e45cafce4123b2 -size 2214065 +oid sha256:43eecd43e6c46a74e6de15982cf36b3fc0e0787ce12ddd061ecf3ee143d3e31c +size 2247094 diff --git a/data/covidtracking/states-metadata.json b/data/covidtracking/states-metadata.json new file mode 100644 index 0000000000000000000000000000000000000000..d1f619dad2776e49242c8f94633f17b7531ec5e5 --- /dev/null +++ b/data/covidtracking/states-metadata.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b48c0fd54e43b491a56fd0b2f36c4d683e742940b5172b07e53b622191b93eee +size 30913 diff --git a/runs/download-covidtracking-data.runs.ipynb b/runs/download-covidtracking-data.runs.ipynb index 4a3e31b8a4177ca4855e8159e29a4fa67434b2ec..b9e8c12bc30ef2e8b7839ef8a323672eaf6008e6 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": 4.308881, - "end_time": "2020-05-12T07:26:15.161447", + "duration": 4.361081, + "end_time": "2020-05-13T09:08:01.086604", "exception": false, - "start_time": "2020-05-12T07:26:10.852566", + "start_time": "2020-05-13T09:07:56.725523", "status": "completed" }, "tags": [] @@ -25,10 +25,10 @@ "execution_count": 2, "metadata": { "papermill": { - "duration": 0.019705, - "end_time": "2020-05-12T07:26:15.195521", + "duration": 0.018596, + "end_time": "2020-05-13T09:08:01.115959", "exception": false, - "start_time": "2020-05-12T07:26:15.175816", + "start_time": "2020-05-13T09:08:01.097363", "status": "completed" }, "tags": [ @@ -46,10 +46,10 @@ "execution_count": 3, "metadata": { "papermill": { - "duration": 0.019669, - "end_time": "2020-05-12T07:26:15.223947", + "duration": 0.022521, + "end_time": "2020-05-13T09:08:01.146502", "exception": false, - "start_time": "2020-05-12T07:26:15.204278", + "start_time": "2020-05-13T09:08:01.123981", "status": "completed" }, "tags": [ @@ -59,7 +59,7 @@ "outputs": [], "source": [ "# Parameters\n", - "PAPERMILL_INPUT_PATH = \"/tmp/osodv0vi/notebooks/process/download-covidtracking-data.ipynb\"\n", + "PAPERMILL_INPUT_PATH = \"/tmp/x8u9p2lc/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": 0.007462, - "end_time": "2020-05-12T07:26:15.240736", + "duration": 0.007505, + "end_time": "2020-05-13T09:08:01.165512", "exception": false, - "start_time": "2020-05-12T07:26:15.233274", + "start_time": "2020-05-13T09:08:01.158007", "status": "completed" }, "tags": [] @@ -87,10 +87,10 @@ "execution_count": 4, "metadata": { "papermill": { - "duration": 0.681241, - "end_time": "2020-05-12T07:26:15.929292", + "duration": 0.732189, + "end_time": "2020-05-13T09:08:01.905618", "exception": false, - "start_time": "2020-05-12T07:26:15.248051", + "start_time": "2020-05-13T09:08:01.173429", "status": "completed" }, "tags": [] @@ -107,10 +107,10 @@ "execution_count": 5, "metadata": { "papermill": { - "duration": 0.026845, - "end_time": "2020-05-12T07:26:15.969206", + "duration": 0.024547, + "end_time": "2020-05-13T09:08:01.945188", "exception": false, - "start_time": "2020-05-12T07:26:15.942361", + "start_time": "2020-05-13T09:08:01.920641", "status": "completed" }, "tags": [] @@ -129,10 +129,10 @@ "execution_count": 6, "metadata": { "papermill": { - "duration": 0.099138, - "end_time": "2020-05-12T07:26:16.078578", + "duration": 0.077147, + "end_time": "2020-05-13T09:08:02.031640", "exception": false, - "start_time": "2020-05-12T07:26:15.979440", + "start_time": "2020-05-13T09:08:01.954493", "status": "completed" }, "tags": [] @@ -246,10 +246,10 @@ "cell_type": "markdown", "metadata": { "papermill": { - "duration": 0.008504, - "end_time": "2020-05-12T07:26:16.096531", + "duration": 0.007787, + "end_time": "2020-05-13T09:08:02.056581", "exception": false, - "start_time": "2020-05-12T07:26:16.088027", + "start_time": "2020-05-13T09:08:02.048794", "status": "completed" }, "tags": [] @@ -263,10 +263,10 @@ "execution_count": 7, "metadata": { "papermill": { - "duration": 4.274671, - "end_time": "2020-05-12T07:26:20.379592", + "duration": 4.091394, + "end_time": "2020-05-13T09:08:06.155844", "exception": false, - "start_time": "2020-05-12T07:26:16.104921", + "start_time": "2020-05-13T09:08:02.064450", "status": "completed" }, "tags": [] @@ -283,10 +283,10 @@ "execution_count": 8, "metadata": { "papermill": { - "duration": 0.030263, - "end_time": "2020-05-12T07:26:20.423251", + "duration": 0.027153, + "end_time": "2020-05-13T09:08:06.195769", "exception": false, - "start_time": "2020-05-12T07:26:20.392988", + "start_time": "2020-05-13T09:08:06.168616", "status": "completed" }, "tags": [] @@ -305,10 +305,10 @@ "execution_count": 9, "metadata": { "papermill": { - "duration": 0.156344, - "end_time": "2020-05-12T07:26:20.592247", + "duration": 0.148828, + "end_time": "2020-05-13T09:08:06.353717", "exception": false, - "start_time": "2020-05-12T07:26:20.435903", + "start_time": "2020-05-13T09:08:06.204889", "status": "completed" }, "tags": [] @@ -318,7 +318,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "3769 data points\n" + "3825 data points\n" ] }, { @@ -368,51 +368,51 @@ " <tbody>\n", " <tr>\n", " <th>0</th>\n", - " <td>20200511</td>\n", + " <td>20200512</td>\n", " <td>AK</td>\n", - " <td>381.0</td>\n", - " <td>28299.0</td>\n", + " <td>383.0</td>\n", + " <td>29578.0</td>\n", " <td>NaN</td>\n", - " <td>7.0</td>\n", + " <td>10.0</td>\n", " <td>NaN</td>\n", " <td>NaN</td>\n", " <td>NaN</td>\n", " <td>NaN</td>\n", " <td>...</td>\n", " <td>NaN</td>\n", - " <td>28680</td>\n", - " <td>28680</td>\n", - " <td>28680</td>\n", + " <td>29961</td>\n", + " <td>29961</td>\n", + " <td>29961</td>\n", " <td>2</td>\n", " <td>0.0</td>\n", " <td>0.0</td>\n", - " <td>1314.0</td>\n", + " <td>1279.0</td>\n", " <td>2.0</td>\n", - " <td>1316.0</td>\n", + " <td>1281.0</td>\n", " </tr>\n", " <tr>\n", " <th>1</th>\n", - " <td>20200511</td>\n", + " <td>20200512</td>\n", " <td>AL</td>\n", - " <td>10009.0</td>\n", - " <td>119435.0</td>\n", + " <td>10310.0</td>\n", + " <td>122908.0</td>\n", " <td>NaN</td>\n", " <td>NaN</td>\n", - " <td>1256.0</td>\n", + " <td>1287.0</td>\n", " <td>NaN</td>\n", - " <td>463.0</td>\n", + " <td>468.0</td>\n", " <td>NaN</td>\n", " <td>...</td>\n", - " <td>1256.0</td>\n", - " <td>129444</td>\n", - " <td>129444</td>\n", - " <td>129444</td>\n", + " <td>1287.0</td>\n", + " <td>133218</td>\n", + " <td>133218</td>\n", + " <td>133218</td>\n", " <td>1</td>\n", - " <td>8.0</td>\n", - " <td>16.0</td>\n", - " <td>1791.0</td>\n", - " <td>232.0</td>\n", - " <td>2023.0</td>\n", + " <td>28.0</td>\n", + " <td>31.0</td>\n", + " <td>3473.0</td>\n", + " <td>301.0</td>\n", + " <td>3774.0</td>\n", " </tr>\n", " </tbody>\n", "</table>\n", @@ -421,24 +421,24 @@ ], "text/plain": [ " date state positive negative pending hospitalizedCurrently \\\n", - "0 20200511 AK 381.0 28299.0 NaN 7.0 \n", - "1 20200511 AL 10009.0 119435.0 NaN NaN \n", + "0 20200512 AK 383.0 29578.0 NaN 10.0 \n", + "1 20200512 AL 10310.0 122908.0 NaN NaN \n", "\n", " hospitalizedCumulative inIcuCurrently inIcuCumulative \\\n", "0 NaN NaN NaN \n", - "1 1256.0 NaN 463.0 \n", + "1 1287.0 NaN 468.0 \n", "\n", " onVentilatorCurrently ... hospitalized total totalTestResults posNeg \\\n", - "0 NaN ... NaN 28680 28680 28680 \n", - "1 NaN ... 1256.0 129444 129444 129444 \n", + "0 NaN ... NaN 29961 29961 29961 \n", + "1 NaN ... 1287.0 133218 133218 133218 \n", "\n", " fips deathIncrease hospitalizedIncrease negativeIncrease \\\n", - "0 2 0.0 0.0 1314.0 \n", - "1 1 8.0 16.0 1791.0 \n", + "0 2 0.0 0.0 1279.0 \n", + "1 1 28.0 31.0 3473.0 \n", "\n", " positiveIncrease totalTestResultsIncrease \n", - "0 2.0 1316.0 \n", - "1 232.0 2023.0 \n", + "0 2.0 1281.0 \n", + "1 301.0 3774.0 \n", "\n", "[2 rows x 27 columns]" ] @@ -474,18 +474,18 @@ "version": "3.7.3" }, "papermill": { - "duration": 11.757658, - "end_time": "2020-05-12T07:26:20.913569", + "duration": 11.412102, + "end_time": "2020-05-13T09:08:06.675940", "environment_variables": {}, "exception": null, - "input_path": "/tmp/osodv0vi/notebooks/process/download-covidtracking-data.ipynb", + "input_path": "/tmp/x8u9p2lc/notebooks/process/download-covidtracking-data.ipynb", "output_path": "runs/download-covidtracking-data.runs.ipynb", "parameters": { - "PAPERMILL_INPUT_PATH": "/tmp/osodv0vi/notebooks/process/download-covidtracking-data.ipynb", + "PAPERMILL_INPUT_PATH": "/tmp/x8u9p2lc/notebooks/process/download-covidtracking-data.ipynb", "PAPERMILL_OUTPUT_PATH": "runs/download-covidtracking-data.runs.ipynb", "out_folder": "data/covidtracking" }, - "start_time": "2020-05-12T07:26:09.155911", + "start_time": "2020-05-13T09:07:55.263838", "version": "1.1.0" } },