From 378f9004cc645ef99b4ccc1c87077529566bbd37 Mon Sep 17 00:00:00 2001
From: "CR (covid cron)" <cramakri+covid-cron@ethz.ch>
Date: Tue, 21 Apr 2020 22:08:13 +0000
Subject: [PATCH] renku rerun data/covidtracking/states-metadata.json
data/covidtracking/states-daily.json
---
.../4cd3481674d6469c84431a2cae839e41.cwl | 120 +++++++++++++++
data/covidtracking/states-daily.json | 4 +-
data/covidtracking/states-metadata.json | 4 +-
runs/download-covidtracking-data.runs.ipynb | 144 +++++++++---------
4 files changed, 196 insertions(+), 76 deletions(-)
create mode 100644 .renku/workflow/4cd3481674d6469c84431a2cae839e41.cwl
diff --git a/.renku/workflow/4cd3481674d6469c84431a2cae839e41.cwl b/.renku/workflow/4cd3481674d6469c84431a2cae839e41.cwl
new file mode 100644
index 000000000..4dfcafb5b
--- /dev/null
+++ b/.renku/workflow/4cd3481674d6469c84431a2cae839e41.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_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-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 0055f84df..23eecf3ed 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 sha256:98f82a5b65d02f0a717ed6030c45888d1684f9a25887ff1b1b3510b3b5e29424
-size 1234525
+oid sha256:aabaf92ddac15e79847cfe725b4ea1dc6f95bae48ab824b4ad3fb3d230f0839c
+size 1296997
diff --git a/data/covidtracking/states-metadata.json b/data/covidtracking/states-metadata.json
index 74d937fcc..91869fdd0 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 sha256:71e330ddff6f2314bcdda80dd2311315c1cad1dcd22dc534c5f35de31dbec8a3
-size 28275
+oid sha256:e09dc92ab8fd8ee1b8f62d98ab273afb13132e32eac757e6ee4128e6c5d975f3
+size 28635
diff --git a/runs/download-covidtracking-data.runs.ipynb b/runs/download-covidtracking-data.runs.ipynb
index 628587f24..e5850d550 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": 0.670971,
- "end_time": "2020-04-20T09:46:45.215508",
+ "duration": 5.532004,
+ "end_time": "2020-04-21T22:08:11.855758",
"exception": false,
- "start_time": "2020-04-20T09:46:44.544537",
+ "start_time": "2020-04-21T22:08:06.323754",
"status": "completed"
},
"tags": []
@@ -25,10 +25,10 @@
"execution_count": 2,
"metadata": {
"papermill": {
- "duration": 0.017354,
- "end_time": "2020-04-20T09:46:45.242288",
+ "duration": 0.018553,
+ "end_time": "2020-04-21T22:08:11.885551",
"exception": false,
- "start_time": "2020-04-20T09:46:45.224934",
+ "start_time": "2020-04-21T22:08:11.866998",
"status": "completed"
},
"tags": [
@@ -46,10 +46,10 @@
"execution_count": 3,
"metadata": {
"papermill": {
- "duration": 0.017449,
- "end_time": "2020-04-20T09:46:45.268250",
+ "duration": 0.020644,
+ "end_time": "2020-04-21T22:08:11.915312",
"exception": false,
- "start_time": "2020-04-20T09:46:45.250801",
+ "start_time": "2020-04-21T22:08:11.894668",
"status": "completed"
},
"tags": [
@@ -59,7 +59,7 @@
"outputs": [],
"source": [
"# Parameters\n",
- "PAPERMILL_INPUT_PATH = \"/tmp/ukjvdt8w/notebooks/process/download-covidtracking-data.ipynb\"\n",
+ "PAPERMILL_INPUT_PATH = \"/tmp/ho22unoi/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": 0.007493,
- "end_time": "2020-04-20T09:46:45.284450",
+ "duration": 0.00694,
+ "end_time": "2020-04-21T22:08:11.931022",
"exception": false,
- "start_time": "2020-04-20T09:46:45.276957",
+ "start_time": "2020-04-21T22:08:11.924082",
"status": "completed"
},
"tags": []
@@ -87,10 +87,10 @@
"execution_count": 4,
"metadata": {
"papermill": {
- "duration": 0.698636,
- "end_time": "2020-04-20T09:46:45.990193",
+ "duration": 0.203156,
+ "end_time": "2020-04-21T22:08:12.141364",
"exception": false,
- "start_time": "2020-04-20T09:46:45.291557",
+ "start_time": "2020-04-21T22:08:11.938208",
"status": "completed"
},
"tags": []
@@ -107,10 +107,10 @@
"execution_count": 5,
"metadata": {
"papermill": {
- "duration": 0.022706,
- "end_time": "2020-04-20T09:46:46.025249",
+ "duration": 0.021073,
+ "end_time": "2020-04-21T22:08:12.174309",
"exception": false,
- "start_time": "2020-04-20T09:46:46.002543",
+ "start_time": "2020-04-21T22:08:12.153236",
"status": "completed"
},
"tags": []
@@ -129,10 +129,10 @@
"execution_count": 6,
"metadata": {
"papermill": {
- "duration": 0.065726,
- "end_time": "2020-04-20T09:46:46.100845",
+ "duration": 0.074361,
+ "end_time": "2020-04-21T22:08:12.258143",
"exception": false,
- "start_time": "2020-04-20T09:46:46.035119",
+ "start_time": "2020-04-21T22:08:12.183782",
"status": "completed"
},
"tags": []
@@ -246,10 +246,10 @@
"cell_type": "markdown",
"metadata": {
"papermill": {
- "duration": 0.008043,
- "end_time": "2020-04-20T09:46:46.120804",
+ "duration": 0.011041,
+ "end_time": "2020-04-21T22:08:12.281752",
"exception": false,
- "start_time": "2020-04-20T09:46:46.112761",
+ "start_time": "2020-04-21T22:08:12.270711",
"status": "completed"
},
"tags": []
@@ -263,10 +263,10 @@
"execution_count": 7,
"metadata": {
"papermill": {
- "duration": 0.16965,
- "end_time": "2020-04-20T09:46:46.298759",
+ "duration": 0.252539,
+ "end_time": "2020-04-21T22:08:12.544413",
"exception": false,
- "start_time": "2020-04-20T09:46:46.129109",
+ "start_time": "2020-04-21T22:08:12.291874",
"status": "completed"
},
"tags": []
@@ -283,10 +283,10 @@
"execution_count": 8,
"metadata": {
"papermill": {
- "duration": 0.027576,
- "end_time": "2020-04-20T09:46:46.338007",
+ "duration": 0.025859,
+ "end_time": "2020-04-21T22:08:12.580466",
"exception": false,
- "start_time": "2020-04-20T09:46:46.310431",
+ "start_time": "2020-04-21T22:08:12.554607",
"status": "completed"
},
"tags": []
@@ -305,10 +305,10 @@
"execution_count": 9,
"metadata": {
"papermill": {
- "duration": 0.116367,
- "end_time": "2020-04-20T09:46:46.465702",
+ "duration": 0.119658,
+ "end_time": "2020-04-21T22:08:12.712624",
"exception": false,
- "start_time": "2020-04-20T09:46:46.349335",
+ "start_time": "2020-04-21T22:08:12.592966",
"status": "completed"
},
"tags": []
@@ -318,7 +318,7 @@
"name": "stdout",
"output_type": "stream",
"text": [
- "2508 data points\n"
+ "2620 data points\n"
]
},
{
@@ -368,51 +368,51 @@
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
- " <td>20200419</td>\n",
+ " <td>20200421</td>\n",
" <td>AK</td>\n",
- " <td>319.0</td>\n",
- " <td>9576.0</td>\n",
+ " <td>329.0</td>\n",
+ " <td>10790.0</td>\n",
" <td>NaN</td>\n",
- " <td>37.0</td>\n",
+ " <td>42.0</td>\n",
" <td>36.0</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>...</td>\n",
" <td>36.0</td>\n",
- " <td>9895.0</td>\n",
- " <td>9895.0</td>\n",
- " <td>9895.0</td>\n",
+ " <td>11119.0</td>\n",
+ " <td>11119.0</td>\n",
+ " <td>11119.0</td>\n",
" <td>2</td>\n",
" <td>0.0</td>\n",
" <td>0.0</td>\n",
- " <td>235.0</td>\n",
- " <td>5.0</td>\n",
- " <td>240.0</td>\n",
+ " <td>987.0</td>\n",
+ " <td>8.0</td>\n",
+ " <td>995.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
- " <td>20200419</td>\n",
+ " <td>20200421</td>\n",
" <td>AL</td>\n",
- " <td>4837.0</td>\n",
- " <td>40875.0</td>\n",
+ " <td>5231.0</td>\n",
+ " <td>43295.0</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
- " <td>641.0</td>\n",
+ " <td>699.0</td>\n",
" <td>NaN</td>\n",
" <td>260.0</td>\n",
" <td>NaN</td>\n",
" <td>...</td>\n",
- " <td>641.0</td>\n",
- " <td>45712.0</td>\n",
- " <td>45712.0</td>\n",
- " <td>45712.0</td>\n",
+ " <td>699.0</td>\n",
+ " <td>48526.0</td>\n",
+ " <td>48526.0</td>\n",
+ " <td>48526.0</td>\n",
" <td>1</td>\n",
- " <td>8.0</td>\n",
- " <td>21.0</td>\n",
- " <td>2992.0</td>\n",
- " <td>182.0</td>\n",
- " <td>3174.0</td>\n",
+ " <td>10.0</td>\n",
+ " <td>58.0</td>\n",
+ " <td>2420.0</td>\n",
+ " <td>206.0</td>\n",
+ " <td>2626.0</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
@@ -421,24 +421,24 @@
],
"text/plain": [
" date state positive negative pending hospitalizedCurrently \\\n",
- "0 20200419 AK 319.0 9576.0 NaN 37.0 \n",
- "1 20200419 AL 4837.0 40875.0 NaN NaN \n",
+ "0 20200421 AK 329.0 10790.0 NaN 42.0 \n",
+ "1 20200421 AL 5231.0 43295.0 NaN NaN \n",
"\n",
" hospitalizedCumulative inIcuCurrently inIcuCumulative \\\n",
"0 36.0 NaN NaN \n",
- "1 641.0 NaN 260.0 \n",
+ "1 699.0 NaN 260.0 \n",
"\n",
" onVentilatorCurrently ... hospitalized total totalTestResults \\\n",
- "0 NaN ... 36.0 9895.0 9895.0 \n",
- "1 NaN ... 641.0 45712.0 45712.0 \n",
+ "0 NaN ... 36.0 11119.0 11119.0 \n",
+ "1 NaN ... 699.0 48526.0 48526.0 \n",
"\n",
" posNeg fips deathIncrease hospitalizedIncrease negativeIncrease \\\n",
- "0 9895.0 2 0.0 0.0 235.0 \n",
- "1 45712.0 1 8.0 21.0 2992.0 \n",
+ "0 11119.0 2 0.0 0.0 987.0 \n",
+ "1 48526.0 1 10.0 58.0 2420.0 \n",
"\n",
" positiveIncrease totalTestResultsIncrease \n",
- "0 5.0 240.0 \n",
- "1 182.0 3174.0 \n",
+ "0 8.0 995.0 \n",
+ "1 206.0 2626.0 \n",
"\n",
"[2 rows x 25 columns]"
]
@@ -474,18 +474,18 @@
"version": "3.7.3"
},
"papermill": {
- "duration": 3.15694,
- "end_time": "2020-04-20T09:46:46.789934",
+ "duration": 8.64246,
+ "end_time": "2020-04-21T22:08:13.031600",
"environment_variables": {},
"exception": null,
- "input_path": "/tmp/ukjvdt8w/notebooks/process/download-covidtracking-data.ipynb",
+ "input_path": "/tmp/ho22unoi/notebooks/process/download-covidtracking-data.ipynb",
"output_path": "runs/download-covidtracking-data.runs.ipynb",
"parameters": {
- "PAPERMILL_INPUT_PATH": "/tmp/ukjvdt8w/notebooks/process/download-covidtracking-data.ipynb",
+ "PAPERMILL_INPUT_PATH": "/tmp/ho22unoi/notebooks/process/download-covidtracking-data.ipynb",
"PAPERMILL_OUTPUT_PATH": "runs/download-covidtracking-data.runs.ipynb",
"out_folder": "data/covidtracking"
},
- "start_time": "2020-04-20T09:46:43.632994",
+ "start_time": "2020-04-21T22:08:04.389140",
"version": "1.1.0"
}
},
--
GitLab