From 812a4a3c6f29e2462c55fff604f5e5b6644d9326 Mon Sep 17 00:00:00 2001 From: "CR (covid cron)" <beepbop@example.com> Date: Fri, 12 Jun 2020 07:28:01 +0000 Subject: [PATCH] renku rerun data/covidtracking/states-metadata.json data/covidtracking/states-daily.json --- .../4d4fd4a4bd4143109c05a9a4cf6100c6.cwl | 120 ++++++++++++++++ data/covidtracking/states-daily.json | 4 +- data/covidtracking/states-metadata.json | 4 +- runs/download-covidtracking-data.runs.ipynb | 128 +++++++++--------- 4 files changed, 188 insertions(+), 68 deletions(-) create mode 100644 .renku/workflow/4d4fd4a4bd4143109c05a9a4cf6100c6.cwl diff --git a/.renku/workflow/4d4fd4a4bd4143109c05a9a4cf6100c6.cwl b/.renku/workflow/4d4fd4a4bd4143109c05a9a4cf6100c6.cwl new file mode 100644 index 000000000..75a9e4293 --- /dev/null +++ b/.renku/workflow/4d4fd4a4bd4143109c05a9a4cf6100c6.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-daily.json + streamable: false + type: string + input_6: + default: states-metadata.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_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-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 + 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-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 + - *id002 + successCodes: [] + temporaryFailCodes: [] diff --git a/data/covidtracking/states-daily.json b/data/covidtracking/states-daily.json index 60fc7bacc..0088ba7b9 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:f267857f071efbd257d39a8e53597a33766547cafe42a33c8fe223186c40a1af -size 4803258 +oid sha256:c094d0a006dab360e7a6e145993c356d376491c000065cb56bcccad90615f88b +size 4853561 diff --git a/data/covidtracking/states-metadata.json b/data/covidtracking/states-metadata.json index 4cdbf31f5..35d1920c5 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:b7f6ed28bdb6086871fec5a39085c151ec8b06e4e74304f4a6dac28966e241d1 -size 37118 +oid sha256:de4dc030edf206a7230a45472b7f29ecc2df492951ebf966a134a316a932b54f +size 37594 diff --git a/runs/download-covidtracking-data.runs.ipynb b/runs/download-covidtracking-data.runs.ipynb index 69e7931ba..5e1967d20 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": 0.487122, - "end_time": "2020-06-11T07:27:32.572137", + "duration": 2.865511, + "end_time": "2020-06-12T07:27:59.346077", "exception": false, - "start_time": "2020-06-11T07:27:32.085015", + "start_time": "2020-06-12T07:27:56.480566", "status": "completed" }, "tags": [] @@ -25,10 +25,10 @@ "execution_count": 2, "metadata": { "papermill": { - "duration": 0.021695, - "end_time": "2020-06-11T07:27:32.609291", + "duration": 0.02443, + "end_time": "2020-06-12T07:27:59.385536", "exception": false, - "start_time": "2020-06-11T07:27:32.587596", + "start_time": "2020-06-12T07:27:59.361106", "status": "completed" }, "tags": [ @@ -46,10 +46,10 @@ "execution_count": 3, "metadata": { "papermill": { - "duration": 0.021591, - "end_time": "2020-06-11T07:27:32.641347", + "duration": 0.023488, + "end_time": "2020-06-12T07:27:59.423778", "exception": false, - "start_time": "2020-06-11T07:27:32.619756", + "start_time": "2020-06-12T07:27:59.400290", "status": "completed" }, "tags": [ @@ -59,7 +59,7 @@ "outputs": [], "source": [ "# Parameters\n", - "PAPERMILL_INPUT_PATH = \"/tmp/iyyjg3ga/notebooks/process/download-covidtracking-data.ipynb\"\n", + "PAPERMILL_INPUT_PATH = \"/tmp/ns7xs7kt/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": 0.006855, - "end_time": "2020-06-11T07:27:32.657727", + "duration": 0.007475, + "end_time": "2020-06-12T07:27:59.442200", "exception": false, - "start_time": "2020-06-11T07:27:32.650872", + "start_time": "2020-06-12T07:27:59.434725", "status": "completed" }, "tags": [] @@ -87,10 +87,10 @@ "execution_count": 4, "metadata": { "papermill": { - "duration": 0.171513, - "end_time": "2020-06-11T07:27:32.835574", + "duration": 0.21893, + "end_time": "2020-06-12T07:27:59.667840", "exception": false, - "start_time": "2020-06-11T07:27:32.664061", + "start_time": "2020-06-12T07:27:59.448910", "status": "completed" }, "tags": [] @@ -107,10 +107,10 @@ "execution_count": 5, "metadata": { "papermill": { - "duration": 0.026162, - "end_time": "2020-06-11T07:27:32.876893", + "duration": 0.025752, + "end_time": "2020-06-12T07:27:59.708723", "exception": false, - "start_time": "2020-06-11T07:27:32.850731", + "start_time": "2020-06-12T07:27:59.682971", "status": "completed" }, "tags": [] @@ -129,10 +129,10 @@ "execution_count": 6, "metadata": { "papermill": { - "duration": 0.074292, - "end_time": "2020-06-11T07:27:32.960997", + "duration": 0.074061, + "end_time": "2020-06-12T07:27:59.794796", "exception": false, - "start_time": "2020-06-11T07:27:32.886705", + "start_time": "2020-06-12T07:27:59.720735", "status": "completed" }, "tags": [] @@ -249,10 +249,10 @@ "cell_type": "markdown", "metadata": { "papermill": { - "duration": 0.007967, - "end_time": "2020-06-11T07:27:32.983385", + "duration": 0.00787, + "end_time": "2020-06-12T07:27:59.816778", "exception": false, - "start_time": "2020-06-11T07:27:32.975418", + "start_time": "2020-06-12T07:27:59.808908", "status": "completed" }, "tags": [] @@ -266,10 +266,10 @@ "execution_count": 7, "metadata": { "papermill": { - "duration": 0.191262, - "end_time": "2020-06-11T07:27:33.181607", + "duration": 0.203847, + "end_time": "2020-06-12T07:28:00.027355", "exception": false, - "start_time": "2020-06-11T07:27:32.990345", + "start_time": "2020-06-12T07:27:59.823508", "status": "completed" }, "tags": [] @@ -286,10 +286,10 @@ "execution_count": 8, "metadata": { "papermill": { - "duration": 0.033195, - "end_time": "2020-06-11T07:27:33.231102", + "duration": 0.033323, + "end_time": "2020-06-12T07:28:00.078483", "exception": false, - "start_time": "2020-06-11T07:27:33.197907", + "start_time": "2020-06-12T07:28:00.045160", "status": "completed" }, "tags": [] @@ -308,10 +308,10 @@ "execution_count": 9, "metadata": { "papermill": { - "duration": 0.210495, - "end_time": "2020-06-11T07:27:33.456670", + "duration": 0.212287, + "end_time": "2020-06-12T07:28:00.306197", "exception": false, - "start_time": "2020-06-11T07:27:33.246175", + "start_time": "2020-06-12T07:28:00.093910", "status": "completed" }, "tags": [] @@ -321,7 +321,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "5449 data points\n" + "5505 data points\n" ] }, { @@ -371,21 +371,21 @@ " <tbody>\n", " <tr>\n", " <th>0</th>\n", - " <td>20200610</td>\n", + " <td>20200611</td>\n", " <td>AK</td>\n", - " <td>593.0</td>\n", - " <td>67127.0</td>\n", + " <td>611.0</td>\n", + " <td>68400.0</td>\n", " <td>NaN</td>\n", - " <td>22.0</td>\n", + " <td>18.0</td>\n", " <td>NaN</td>\n", " <td>NaN</td>\n", " <td>NaN</td>\n", - " <td>3.0</td>\n", + " <td>4.0</td>\n", " <td>...</td>\n", - " <td>67720</td>\n", + " <td>69011</td>\n", " <td>0</td>\n", " <td>0</td>\n", - " <td>c9486e6829e118950bb3df693089884e90df1761</td>\n", + " <td>81be29c53674efe5eb683d80c80ed7b7a012f113</td>\n", " <td>0</td>\n", " <td>0</td>\n", " <td>0</td>\n", @@ -395,21 +395,21 @@ " </tr>\n", " <tr>\n", " <th>1</th>\n", - " <td>20200610</td>\n", + " <td>20200611</td>\n", " <td>AL</td>\n", - " <td>21989.0</td>\n", - " <td>251068.0</td>\n", + " <td>22845.0</td>\n", + " <td>255981.0</td>\n", " <td>NaN</td>\n", - " <td>629.0</td>\n", - " <td>2129.0</td>\n", + " <td>647.0</td>\n", + " <td>2165.0</td>\n", " <td>NaN</td>\n", - " <td>646.0</td>\n", + " <td>650.0</td>\n", " <td>NaN</td>\n", " <td>...</td>\n", - " <td>273057</td>\n", - " <td>15</td>\n", - " <td>42</td>\n", - " <td>19414d6567b3415847b4d4d5d4c191f2b954647f</td>\n", + " <td>278826</td>\n", + " <td>11</td>\n", + " <td>36</td>\n", + " <td>14deff225758b20ec967e765c4fb230342d28c70</td>\n", " <td>0</td>\n", " <td>0</td>\n", " <td>0</td>\n", @@ -424,20 +424,20 @@ ], "text/plain": [ " date state positive negative pending hospitalizedCurrently \\\n", - "0 20200610 AK 593.0 67127.0 NaN 22.0 \n", - "1 20200610 AL 21989.0 251068.0 NaN 629.0 \n", + "0 20200611 AK 611.0 68400.0 NaN 18.0 \n", + "1 20200611 AL 22845.0 255981.0 NaN 647.0 \n", "\n", " hospitalizedCumulative inIcuCurrently inIcuCumulative \\\n", "0 NaN NaN NaN \n", - "1 2129.0 NaN 646.0 \n", + "1 2165.0 NaN 650.0 \n", "\n", " onVentilatorCurrently ... posNeg deathIncrease hospitalizedIncrease \\\n", - "0 3.0 ... 67720 0 0 \n", - "1 NaN ... 273057 15 42 \n", + "0 4.0 ... 69011 0 0 \n", + "1 NaN ... 278826 11 36 \n", "\n", " hash commercialScore \\\n", - "0 c9486e6829e118950bb3df693089884e90df1761 0 \n", - "1 19414d6567b3415847b4d4d5d4c191f2b954647f 0 \n", + "0 81be29c53674efe5eb683d80c80ed7b7a012f113 0 \n", + "1 14deff225758b20ec967e765c4fb230342d28c70 0 \n", "\n", " negativeRegularScore negativeScore positiveScore score grade \n", "0 0 0 0 0 \n", @@ -477,18 +477,18 @@ "version": "3.7.3" }, "papermill": { - "duration": 2.546059, - "end_time": "2020-06-11T07:27:33.780969", + "duration": 5.690537, + "end_time": "2020-06-12T07:28:00.738243", "environment_variables": {}, "exception": null, - "input_path": "/tmp/iyyjg3ga/notebooks/process/download-covidtracking-data.ipynb", + "input_path": "/tmp/ns7xs7kt/notebooks/process/download-covidtracking-data.ipynb", "output_path": "runs/download-covidtracking-data.runs.ipynb", "parameters": { - "PAPERMILL_INPUT_PATH": "/tmp/iyyjg3ga/notebooks/process/download-covidtracking-data.ipynb", + "PAPERMILL_INPUT_PATH": "/tmp/ns7xs7kt/notebooks/process/download-covidtracking-data.ipynb", "PAPERMILL_OUTPUT_PATH": "runs/download-covidtracking-data.runs.ipynb", "out_folder": "data/covidtracking" }, - "start_time": "2020-06-11T07:27:31.234910", + "start_time": "2020-06-12T07:27:55.047706", "version": "1.1.0" } }, -- GitLab