Skip to content

Instantly share code, notes, and snippets.

@pbugnion

pbugnion/01_readme.md

Last active September 14, 2021 05:36
Show Gist options

  • Select an option

    Learn more about clone URLs
  • Save pbugnion/a4a24d6f54fe9f6239aae90569197e9e to your computer and use it in GitHub Desktop.

Select an option

Learn more about clone URLs
Save pbugnion/a4a24d6f54fe9f6239aae90569197e9e to your computer and use it in GitHub Desktop.
Download ZIP

Revisions

  1. pbugnion revised this gist Jun 5, 2018. 1 changed file with 4 additions and 2 deletions.
    6 changes: 4 additions & 2 deletions 01_readme.md
    View file
    Original file line number Diff line number Diff line change
    @@ -28,7 +28,7 @@ pip install apache-airflow

    mkdir -p $AIRFLOW_HOME

    wget --output-document /project/airflow/airflow.cfg $SOURCE/a0ec0a53459bbfd52a61c4eb766cac22c79d5f3e/airflow.cfg
    wget --output-document $AIRFLOW_HOME/airflow.cfg $SOURCE/airflow.cfg

    echo "export AIRFLOW_HOME=$AIRFLOW_HOME" | sudo tee /etc/profile.d/airflow.sh

    @@ -46,6 +46,8 @@ script:
    AIRFLOW_HOME=/project/airflow
    SOURCE=https://gist.github.com/pbugnion/a4a24d6f54fe9f6239aae90569197e9e/raw
    echo "export AIRFLOW_HOME=$AIRFLOW_HOME" | sudo tee /etc/profile.d/airflow.sh
    sudo sv stop jupyter
    wget --output-document /tmp/airflow-run $SOURCE/run
    @@ -84,7 +86,7 @@ wget --output-document /project/stock-price/fetch_stock.py $SOURCE/fetch_stock.p
    wget --output-document /project/stock-price/publish_report.py $SOURCE/publish_report.py
    wget --output-document /project/stock-price/plot-stock-price.ipynb $SOURCE/plot-stock-price.ipynb

    wget --output-document /project/airflow/stock_price.py $SOURCE/stock_price_dag.py
    wget --output-document /project/airflow/dags/stock_price.py $SOURCE/stock_price_dag.py
    ```

    Then create a report by publishing the `plot-stock-price` notebook in
  2. pbugnion revised this gist Jun 5, 2018. 1 changed file with 1 addition and 0 deletions.
    1 change: 1 addition & 0 deletions 01_readme.md
    View file
    Original file line number Diff line number Diff line change
    @@ -82,6 +82,7 @@ mkdir -p /project/stock-price /project/stock-price/data

    wget --output-document /project/stock-price/fetch_stock.py $SOURCE/fetch_stock.py
    wget --output-document /project/stock-price/publish_report.py $SOURCE/publish_report.py
    wget --output-document /project/stock-price/plot-stock-price.ipynb $SOURCE/plot-stock-price.ipynb

    wget --output-document /project/airflow/stock_price.py $SOURCE/stock_price_dag.py
    ```
  3. pbugnion revised this gist Jun 5, 2018. 1 changed file with 1 addition and 1 deletion.
    2 changes: 1 addition & 1 deletion 01_readme.md
    View file
    Original file line number Diff line number Diff line change
    @@ -30,7 +30,7 @@ mkdir -p $AIRFLOW_HOME

    wget --output-document /project/airflow/airflow.cfg $SOURCE/a0ec0a53459bbfd52a61c4eb766cac22c79d5f3e/airflow.cfg

    echo "export AIRFLOW_HOME=$AIRFLOW_HOME" > /etc/profile.d/airflow.sh
    echo "export AIRFLOW_HOME=$AIRFLOW_HOME" | sudo tee /etc/profile.d/airflow.sh

    airflow initdb
    ```
  4. pbugnion revised this gist Jun 5, 2018. 1 changed file with 4 additions and 0 deletions.
    4 changes: 4 additions & 0 deletions 01_readme.md
    View file
    Original file line number Diff line number Diff line change
    @@ -91,6 +91,10 @@ Then create a report by publishing the `plot-stock-price` notebook in
    UUID in the URL when you are viewing the report). Edit the DAG definition
    with the correct UUID.

    The file `/project/airflow/dags/stock_price.py` is the airflow DAG
    definition: it defines what tasks are available in the DAG, and how
    they depend on each other.

    This pipeline is scheduled to run every day at midnight. You can also
    run the pipeline for a particular day or set of days by running the
    following command in a *new* terminal on the instance running the
  5. pbugnion revised this gist Jun 5, 2018. 3 changed files with 22 additions and 4 deletions.
    7 changes: 7 additions & 0 deletions 01_readme.md
    View file
    Original file line number Diff line number Diff line change
    @@ -86,6 +86,11 @@ wget --output-document /project/stock-price/publish_report.py $SOURCE/publish_re
    wget --output-document /project/airflow/stock_price.py $SOURCE/stock_price_dag.py
    ```

    Then create a report by publishing the `plot-stock-price` notebook in
    `/project/stock-price`. Get the report ID for that report (the last
    UUID in the URL when you are viewing the report). Edit the DAG definition
    with the correct UUID.

    This pipeline is scheduled to run every day at midnight. You can also
    run the pipeline for a particular day or set of days by running the
    following command in a *new* terminal on the instance running the
    @@ -97,3 +102,5 @@ airflow backfill stock_price -s 2018-03-01 -e 2018-03-06

    This will fetch the stock price for dates between the 1st and the 6th
    of March 2018.

    ![](https://gist.github.com/pbugnion/a4a24d6f54fe9f6239aae90569197e9e/raw/report-example.png)
    Binary file added report-example.png
    View file
    Loading
    Sorry, something went wrong. Reload?
    Sorry, we cannot display this file.
    Sorry, this file is invalid so it cannot be displayed.
    19 changes: 15 additions & 4 deletions stock_price_dag.py
    View file
    Original file line number Diff line number Diff line change
    @@ -2,6 +2,8 @@
    from airflow.operators.bash_operator import BashOperator
    from datetime import datetime, timedelta

    REPORT_ID = 'b0b5e1c5-d7c9-4ae5-8b02-ce0f4a7ddaee'

    default_args = {
    'owner': 'airflow',
    'depends_on_past': False,
    @@ -19,18 +21,27 @@
    )


    fetch_command = '/opt/anaconda/envs/Python3/bin/python /project/stock-price/fetch_stock.py GOOG {{ ds }} {{ ds }}'
    fetch_command = (
    '/opt/anaconda/envs/Python3/bin/python '
    '/project/stock-price/fetch_stock.py GOOG {{ ds }} {{ ds }}'
    )

    fetch_operator = BashOperator(
    task_id='fetch_price',
    bash_command=fetch_command,
    dag = dag
    dag=dag
    )

    publish_operator = BashOperator(
    task_id='publish',
    bash_command='/opt/anaconda/envs/Python3/bin/python /project/stock-price/publish_report.py',
    bash_command=(
    '/opt/anaconda/envs/Python3/bin/python '
    '/project/stock-price/publish_report.py '
    f'{REPORT_ID} '
    '/project/stock-price/plot-stock-price.ipynb'
    ),
    dag=dag
    )

    publish_operator.set_upstream(fetch_operator)

    publish_operator.set_upstream(fetch_operator)
  6. pbugnion revised this gist Jun 5, 2018. 1 changed file with 3 additions and 2 deletions.
    5 changes: 3 additions & 2 deletions publish_report.py
    View file
    Original file line number Diff line number Diff line change
    @@ -7,10 +7,11 @@
    from nbconvert.preprocessors import ExecutePreprocessor

    import sml.auth
    import sml.config


    HUDSON_URL = 'https://hudson.platform.asidata.science'
    TAVERN_URL = 'https://tavern.platform.asidata.science'
    HUDSON_URL = sml.config.url_for_service('hudson')
    TAVERN_URL = sml.config.url_for_service('tavern')


    class InputError(Exception):
  7. pbugnion revised this gist Jun 5, 2018. 1 changed file with 20 additions and 7 deletions.
    27 changes: 20 additions & 7 deletions publish_report.py
    View file
    Original file line number Diff line number Diff line change
    @@ -1,18 +1,17 @@
    import os
    import argparse

    import requests
    import sherlockml.auth

    import nbformat
    from nbconvert.preprocessors import ExecutePreprocessor

    import sml.auth


    HUDSON_URL = 'https://hudson.platform.asidata.science'
    TAVERN_URL = 'https://tavern.platform.asidata.science'

    REPORT_ID = 'b0b5e1c5-d7c9-4ae5-8b02-ce0f4a7ddaee'
    NOTEBOOK_PATH = '/project/stock-price/plot-stock-price.ipynb'


    class InputError(Exception):
    pass
    @@ -58,7 +57,7 @@ def run(report_id, notebook, run_notebook=True):
    reconstructed_notebook_path = '/'.join(reconstructed_notebook_path)

    # Authenticate and publish
    auth_headers = sherlockml.auth.auth_headers()
    auth_headers = sml.auth.auth_headers()
    user_id_response = requests.get(
    HUDSON_URL + '/authenticate',
    headers=auth_headers
    @@ -75,6 +74,20 @@ def run(report_id, notebook, run_notebook=True):
    json=body,
    headers=auth_headers)
    create_version_response.raise_for_status()


    run(REPORT_ID, NOTEBOOK_PATH)

    def parse_command_line():
    parser = argparse.ArgumentParser(
    description='Run and publish a notebook')
    parser.add_argument('report_id', help='ID of the report to update')
    parser.add_argument(
    'notebook_path',
    help='Path of the notebook to publish'
    )
    args = parser.parse_args()
    return (args.report_id, args.notebook_path)


    if __name__ == '__main__':
    report_id, notebook_path = parse_command_line()
    run(report_id, notebook_path)
  8. pbugnion revised this gist Jun 5, 2018. 1 changed file with 4 additions and 4 deletions.
    8 changes: 4 additions & 4 deletions 01_readme.md
    View file
    Original file line number Diff line number Diff line change
    @@ -86,14 +86,14 @@ wget --output-document /project/stock-price/publish_report.py $SOURCE/publish_re
    wget --output-document /project/airflow/stock_price.py $SOURCE/stock_price_dag.py
    ```

    You can then run the pipeline by running the following command in a
    *new* terminal on the instance running the *airflow* webserver:
    This pipeline is scheduled to run every day at midnight. You can also
    run the pipeline for a particular day or set of days by running the
    following command in a *new* terminal on the instance running the
    *airflow* webserver:

    ```
    airflow backfill stock_price -s 2018-03-01 -e 2018-03-06
    ```

    This will fetch the stock price for dates between the 1st and the 6th
    of March 2018.

    This pipeline is scheduled to run every day at midnight.
  9. pbugnion revised this gist Jun 5, 2018. 1 changed file with 25 additions and 12 deletions.
    37 changes: 25 additions & 12 deletions 01_readme.md
    View file
    Original file line number Diff line number Diff line change
    @@ -1,16 +1,23 @@
    # Airflow on SherlockML

    [Apache Airflow](https://airflow.apache.org/) is an open source tool for creating
    task pipelines. It lets you define sets of tasks and dependencies between
    those tasks, and then takes care of the execution.
    [Apache Airflow](https://airflow.apache.org/) is an open source tool
    for creating task pipelines. It lets you define sets of tasks and
    dependencies between those tasks, and then takes care of the
    execution.

    ![](https://gist.github.com/pbugnion/a4a24d6f54fe9f6239aae90569197e9e/raw/airflow-example.png)

    Airflow can be a useful add-on to SherlockML: you can schedule
    periodic jobs that automatically fetch data or retrain a model or
    publish a report. By defining dependencies between those, you can
    have a resilient, automated data processing pipeline.

    ## Installing Airflow in SherlockML

    ### Setting up a project

    To set up a project, create an environment called `airflow-setup-project` with the following entry in the scripts tab:
    To set up a project, create an environment called
    `airflow-setup-project` with the following entry in the scripts tab:

    ```bash
    AIRFLOW_HOME=/project/airflow
    @@ -32,7 +39,8 @@ Then apply this environment to any Jupyter server within the project.

    ### Running the Airflow webserver

    Create an environment called `airflow-webserver` with a single shell script:
    Create an environment called `airflow-webserver` with a single shell
    script:

    ```
    AIRFLOW_HOME=/project/airflow
    @@ -53,13 +61,17 @@ sudo sv start airflow
    sleep 20 # Wait for airflow to start
    ```

    When this environment is applied to a server in the workspace, it will stop the Jupyter server and replace it with Airflow. You can then click on the server name to open the Airflow dashboard.
    When this environment is applied to a server in the workspace, it will
    stop the Jupyter server and replace it with Airflow. You can then
    click on the server name to open the Airflow dashboard.

    ### Install examples

    This gist contains an example with an airflow pipeline that fetches the Google stock price for a given date,
    saves it in a CSV in the SherlockML workspace (we could also save it in, say, InfluxDB) and
    publishes a SherlockML report with a graph of the stock price over time, including that date.
    This gist contains an example with an airflow pipeline that fetches
    the Google stock price for a given date, saves it in a CSV in the
    SherlockML workspace (we could also save it in, say, InfluxDB) and
    publishes a SherlockML report with a graph of the stock price over
    time, including that date.

    To install the demo, run the following commands in a terminal:

    @@ -74,13 +86,14 @@ wget --output-document /project/stock-price/publish_report.py $SOURCE/publish_re
    wget --output-document /project/airflow/stock_price.py $SOURCE/stock_price_dag.py
    ```

    You can then run the pipeline by running the following command in a *new* terminal on the instance running
    the *airflow* webserver:
    You can then run the pipeline by running the following command in a
    *new* terminal on the instance running the *airflow* webserver:

    ```
    airflow backfill stock_price -s 2018-03-01 -e 2018-03-06
    ```

    This will fetch the stock price for dates between the 1st and the 6th of March 2018.
    This will fetch the stock price for dates between the 1st and the 6th
    of March 2018.

    This pipeline is scheduled to run every day at midnight.
  10. pbugnion revised this gist Jun 5, 2018. 1 changed file with 1 addition and 1 deletion.
    2 changes: 1 addition & 1 deletion 01_readme.md
    View file
    Original file line number Diff line number Diff line change
    @@ -4,7 +4,7 @@
    task pipelines. It lets you define sets of tasks and dependencies between
    those tasks, and then takes care of the execution.

    ![](airflow-example.png)
    ![](https://gist.github.com/pbugnion/a4a24d6f54fe9f6239aae90569197e9e/raw/airflow-example.png)

    ## Installing Airflow in SherlockML

  11. pbugnion revised this gist Jun 5, 2018. 2 changed files with 3 additions and 1 deletion.
    4 changes: 3 additions & 1 deletion 01_readme.md
    View file
    Original file line number Diff line number Diff line change
    @@ -4,6 +4,8 @@
    task pipelines. It lets you define sets of tasks and dependencies between
    those tasks, and then takes care of the execution.

    ![](airflow-example.png)

    ## Installing Airflow in SherlockML

    ### Setting up a project
    @@ -81,4 +83,4 @@ airflow backfill stock_price -s 2018-03-01 -e 2018-03-06

    This will fetch the stock price for dates between the 1st and the 6th of March 2018.

    This pipeline is scheduled to run every day at midnight.
    This pipeline is scheduled to run every day at midnight.
    Binary file added airflow-example.png
    View file
    Loading
    Sorry, something went wrong. Reload?
    Sorry, we cannot display this file.
    Sorry, this file is invalid so it cannot be displayed.
  12. pbugnion revised this gist Jun 5, 2018. No changes.
  13. pbugnion revised this gist Jun 5, 2018. 1 changed file with 3 additions and 1 deletion.
    4 changes: 3 additions & 1 deletion 01_readme.md
    View file
    Original file line number Diff line number Diff line change
    @@ -79,4 +79,6 @@ the *airflow* webserver:
    airflow backfill stock_price -s 2018-03-01 -e 2018-03-06
    ```

    This will fetch the stock price for dates between the 1st and the 6th of March 2018.
    This will fetch the stock price for dates between the 1st and the 6th of March 2018.

    This pipeline is scheduled to run every day at midnight.
  14. pbugnion revised this gist Jun 5, 2018. 1 changed file with 5 additions and 5 deletions.
    10 changes: 5 additions & 5 deletions plot-stock-price.ipynb
    View file
    Original file line number Diff line number Diff line change
    @@ -19,7 +19,7 @@
    "outputs": [],
    "source": [
    "stock_price = 'GOOG'\n",
    "datastore = Path('/project/data')"
    "datastore = Path('/project/stock-price/data')"
    ]
    },
    {
    @@ -31,8 +31,8 @@
    "name": "stdout",
    "output_type": "stream",
    "text": [
    "start date: 2018-01-04 00:00:00\n",
    "end date: 2018-02-23 00:00:00\n"
    "start date: 2018-01-02 00:00:00\n",
    "end date: 2018-01-10 00:00:00\n"
    ]
    }
    ],
    @@ -53,9 +53,9 @@
    "outputs": [
    {
    "data": {
    "image/png": "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\n",
    "image/png": "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\n",
    "text/plain": [
    "<matplotlib.figure.Figure at 0x7f47f6df7128>"
    "<matplotlib.figure.Figure at 0x7f37eb86b358>"
    ]
    },
    "metadata": {},
  15. pbugnion revised this gist Jun 5, 2018. 1 changed file with 1 addition and 1 deletion.
    2 changes: 1 addition & 1 deletion stock_price_dag.py
    View file
    Original file line number Diff line number Diff line change
    @@ -29,7 +29,7 @@

    publish_operator = BashOperator(
    task_id='publish',
    bash_command='/opt/anaconda/envs/Python3/bin/python /project/stock-price/publish.py',
    bash_command='/opt/anaconda/envs/Python3/bin/python /project/stock-price/publish_report.py',
    dag=dag
    )

  16. pbugnion revised this gist Jun 5, 2018. 1 changed file with 1 addition and 1 deletion.
    2 changes: 1 addition & 1 deletion stock_price_dag.py
    View file
    Original file line number Diff line number Diff line change
    @@ -10,7 +10,7 @@
    'email_on_retry': False,
    'retries': 1,
    'retry_delay': timedelta(minutes=5),
    'schedule': '@daily'
    'schedule_interval': '@daily'
    }

    dag = DAG(
  17. pbugnion revised this gist Jun 5, 2018. 2 changed files with 6 additions and 7 deletions.
    10 changes: 4 additions & 6 deletions 01_readme.md
    View file
    Original file line number Diff line number Diff line change
    @@ -36,21 +36,19 @@ Create an environment called `airflow-webserver` with a single shell script:
    AIRFLOW_HOME=/project/airflow
    SOURCE=https://gist.github.com/pbugnion/a4a24d6f54fe9f6239aae90569197e9e/raw
    mkdir -p $AIRFLOW_HOME
    wget --output-document /project/airflow/airflow.cfg $SOURCE/airflow.cfg
    echo "export AIRFLOW_HOME=$AIRFLOW_HOME" | sudo tee /etc/profile.d/airflow.sh
    sudo sv stop jupyter
    wget --output-document /tmp/airflow-run $SOURCE/run
    sudo mkdir -p /etc/service/airflow
    sudo mv /tmp/airflow-run /etc/service/airflow/run
    sudo chown root:root /etc/service/airflow/run
    sudo chmod a+x /etc/service/airflow/run
    sleep 7 # Wait for runit to pick up new service
    sudo sv start airflow
    sleep 20 # Wait for airflow to start
    ```

    When this environment is applied to a server in the workspace, it will stop the Jupyter server and replace it with Airflow. You can then click on the server name to open the Airflow dashboard.
    3 changes: 2 additions & 1 deletion stock_price_dag.py
    View file
    Original file line number Diff line number Diff line change
    @@ -9,7 +9,8 @@
    'email_on_failure': False,
    'email_on_retry': False,
    'retries': 1,
    'retry_delay': timedelta(minutes=5)
    'retry_delay': timedelta(minutes=5),
    'schedule': '@daily'
    }

    dag = DAG(
  18. pbugnion revised this gist Jun 5, 2018. No changes.
  19. pbugnion revised this gist Jun 5, 2018. 2 changed files with 85 additions and 2 deletions.
    85 changes: 84 additions & 1 deletion 01_readme.md
    View file
    Original file line number Diff line number Diff line change
    @@ -1 +1,84 @@
    # Airflow on SherlockML
    # Airflow on SherlockML

    [Apache Airflow](https://airflow.apache.org/) is an open source tool for creating
    task pipelines. It lets you define sets of tasks and dependencies between
    those tasks, and then takes care of the execution.

    ## Installing Airflow in SherlockML

    ### Setting up a project

    To set up a project, create an environment called `airflow-setup-project` with the following entry in the scripts tab:

    ```bash
    AIRFLOW_HOME=/project/airflow
    SOURCE=https://gist.github.com/pbugnion/a4a24d6f54fe9f6239aae90569197e9e/raw/

    source activate Python3
    pip install apache-airflow

    mkdir -p $AIRFLOW_HOME

    wget --output-document /project/airflow/airflow.cfg $SOURCE/a0ec0a53459bbfd52a61c4eb766cac22c79d5f3e/airflow.cfg

    echo "export AIRFLOW_HOME=$AIRFLOW_HOME" > /etc/profile.d/airflow.sh

    airflow initdb
    ```

    Then apply this environment to any Jupyter server within the project.

    ### Running the Airflow webserver

    Create an environment called `airflow-webserver` with a single shell script:

    ```
    AIRFLOW_HOME=/project/airflow
    SOURCE=https://gist.github.com/pbugnion/a4a24d6f54fe9f6239aae90569197e9e/raw
    mkdir -p $AIRFLOW_HOME
    wget --output-document /project/airflow/airflow.cfg $SOURCE/airflow.cfg
    echo "export AIRFLOW_HOME=$AIRFLOW_HOME" | sudo tee /etc/profile.d/airflow.sh
    sudo sv stop jupyter
    wget --output-document /tmp/airflow-run $SOURCE/run
    sudo mkdir -p /etc/service/airflow
    sudo mv /tmp/airflow-run /etc/service/airflow/run
    sleep 7 # Wait for runit to pick up new service
    sudo sv start airflow
    ```

    When this environment is applied to a server in the workspace, it will stop the Jupyter server and replace it with Airflow. You can then click on the server name to open the Airflow dashboard.

    ### Install examples

    This gist contains an example with an airflow pipeline that fetches the Google stock price for a given date,
    saves it in a CSV in the SherlockML workspace (we could also save it in, say, InfluxDB) and
    publishes a SherlockML report with a graph of the stock price over time, including that date.

    To install the demo, run the following commands in a terminal:

    ```bash
    SOURCE=https://gist.github.com/pbugnion/a4a24d6f54fe9f6239aae90569197e9e/raw

    mkdir -p /project/stock-price /project/stock-price/data

    wget --output-document /project/stock-price/fetch_stock.py $SOURCE/fetch_stock.py
    wget --output-document /project/stock-price/publish_report.py $SOURCE/publish_report.py

    wget --output-document /project/airflow/stock_price.py $SOURCE/stock_price_dag.py
    ```

    You can then run the pipeline by running the following command in a *new* terminal on the instance running
    the *airflow* webserver:

    ```
    airflow backfill stock_price -s 2018-03-01 -e 2018-03-06
    ```

    This will fetch the stock price for dates between the 1st and the 6th of March 2018.
    2 changes: 1 addition & 1 deletion fetch_stock.py
    View file
    Original file line number Diff line number Diff line change
    @@ -10,7 +10,7 @@
    logging.basicConfig(level=logging.INFO)

    QUANDL_URL = 'https://www.quandl.com/api/v3/datasets/WIKI/{}.json'
    DATASTORE = Path('/project/data')
    DATASTORE = Path('/project/stock-price/data')

    Configuration = namedtuple(
    'Configuration',
  20. pbugnion revised this gist Jun 5, 2018. 1 changed file with 101 additions and 0 deletions.
    101 changes: 101 additions & 0 deletions plot-stock-price.ipynb
    View file
    Original file line number Diff line number Diff line change
    @@ -0,0 +1,101 @@
    {
    "cells": [
    {
    "cell_type": "code",
    "execution_count": 1,
    "metadata": {},
    "outputs": [],
    "source": [
    "from pathlib import Path\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd"
    ]
    },
    {
    "cell_type": "code",
    "execution_count": 2,
    "metadata": {},
    "outputs": [],
    "source": [
    "stock_price = 'GOOG'\n",
    "datastore = Path('/project/data')"
    ]
    },
    {
    "cell_type": "code",
    "execution_count": 3,
    "metadata": {},
    "outputs": [
    {
    "name": "stdout",
    "output_type": "stream",
    "text": [
    "start date: 2018-01-04 00:00:00\n",
    "end date: 2018-02-23 00:00:00\n"
    ]
    }
    ],
    "source": [
    "with open(datastore / f'{stock_price}.csv') as f:\n",
    " df = pd.read_csv(\n",
    " f, parse_dates=['date'], infer_datetime_format=True\n",
    " )\n",
    "\n",
    "print(f'start date: {min(df.date)}')\n",
    "print(f'end date: {max(df.date)}')"
    ]
    },
    {
    "cell_type": "code",
    "execution_count": 4,
    "metadata": {},
    "outputs": [
    {
    "data": {
    "image/png": "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\n",
    "text/plain": [
    "<matplotlib.figure.Figure at 0x7f47f6df7128>"
    ]
    },
    "metadata": {},
    "output_type": "display_data"
    }
    ],
    "source": [
    "plt.figure(figsize=(8, 6))\n",
    "plt.xlabel('date')\n",
    "plt.ylabel('stock price')\n",
    "plt.plot(df.date, df.price);"
    ]
    },
    {
    "cell_type": "code",
    "execution_count": 5,
    "metadata": {},
    "outputs": [],
    "source": []
    }
    ],
    "metadata": {
    "kernelspec": {
    "display_name": "Python [conda env:Python3]",
    "language": "python",
    "name": "conda-env-Python3-py"
    },
    "language_info": {
    "codemirror_mode": {
    "name": "ipython",
    "version": 3
    },
    "file_extension": ".py",
    "mimetype": "text/x-python",
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
    "version": "3.6.4"
    }
    },
    "nbformat": 4,
    "nbformat_minor": 2
    }
  21. pbugnion revised this gist Jun 5, 2018. 1 changed file with 12 additions and 0 deletions.
    12 changes: 12 additions & 0 deletions run
    View file
    Original file line number Diff line number Diff line change
    @@ -0,0 +1,12 @@
    #!/bin/sh

    # Runit unit file for Airflow webserver

    set -e

    cd "$PROJECT_HOME"

    exec /sbin/setuser sherlock env \
    AIRFLOW_HOME=/project/airflow \
    PATH=/opt/anaconda/envs/Python3/bin:opt/anaconda/bin:"$PATH" \
    airflow webserver
  22. pbugnion revised this gist Jun 5, 2018. 4 changed files with 572 additions and 0 deletions.
    353 changes: 353 additions & 0 deletions airflow.cfg
    View file
    Original file line number Diff line number Diff line change
    @@ -0,0 +1,353 @@
    # Airflow configuration file
    # /project/airflow/airflow.cfg

    [core]
    # The home folder for airflow, default is ~/airflow
    airflow_home = /project/airflow

    # The folder where your airflow pipelines live, most likely a
    # subfolder in a code repository
    # This path must be absolute
    dags_folder = /project/airflow/dags

    # The folder where airflow should store its log files
    # This path must be absolute
    base_log_folder = /project/airflow/logs

    # Airflow can store logs remotely in AWS S3 or Google Cloud Storage. Users
    # must supply a remote location URL (starting with either 's3://...' or
    # 'gs://...') and an Airflow connection id that provides access to the storage
    # location.
    remote_base_log_folder =
    remote_log_conn_id =
    # Use server-side encryption for logs stored in S3
    encrypt_s3_logs = False
    # DEPRECATED option for remote log storage, use remote_base_log_folder instead!
    s3_log_folder =

    # The executor class that airflow should use. Choices include
    # SequentialExecutor, LocalExecutor, CeleryExecutor
    executor = SequentialExecutor

    # The SqlAlchemy connection string to the metadata database.
    # SqlAlchemy supports many different database engine, more information
    # their website
    sql_alchemy_conn = sqlite:////project/airflow/airflow.db

    # The SqlAlchemy pool size is the maximum number of database connections
    # in the pool.
    sql_alchemy_pool_size = 5

    # The SqlAlchemy pool recycle is the number of seconds a connection
    # can be idle in the pool before it is invalidated. This config does
    # not apply to sqlite.
    sql_alchemy_pool_recycle = 3600

    # The amount of parallelism as a setting to the executor. This defines
    # the max number of task instances that should run simultaneously
    # on this airflow installation
    parallelism = 8

    # The number of task instances allowed to run concurrently by the scheduler
    dag_concurrency = 16

    # Are DAGs paused by default at creation
    dags_are_paused_at_creation = True

    # When not using pools, tasks are run in the "default pool",
    # whose size is guided by this config element
    non_pooled_task_slot_count = 128

    # The maximum number of active DAG runs per DAG
    max_active_runs_per_dag = 16

    # Whether to load the examples that ship with Airflow. It's good to
    # get started, but you probably want to set this to False in a production
    # environment
    load_examples = False

    # Where your Airflow plugins are stored
    plugins_folder = /home/sherlock/workspace/airflow/plugins

    # Secret key to save connection passwords in the db
    fernet_key = IrY20uqM6R-OGwrRaWFh4rgdh5hD0n2p7a5jgd-nhxw=

    # Whether to disable pickling dags
    donot_pickle = False

    # How long before timing out a python file import while filling the DagBag
    dagbag_import_timeout = 30

    # The class to use for running task instances in a subprocess
    task_runner = BashTaskRunner

    # If set, tasks without a `run_as_user` argument will be run with this user
    # Can be used to de-elevate a sudo user running Airflow when executing tasks
    default_impersonation =

    # What security module to use (for example kerberos):
    security =

    # Turn unit test mode on (overwrites many configuration options with test
    # values at runtime)
    unit_test_mode = False

    [cli]
    # In what way should the cli access the API. The LocalClient will use the
    # database directly, while the json_client will use the api running on the
    # webserver
    api_client = airflow.api.client.local_client
    endpoint_url = http://localhost:8888

    [api]
    # How to authenticate users of the API
    auth_backend = airflow.api.auth.backend.default

    [operators]
    # The default owner assigned to each new operator, unless
    # provided explicitly or passed via `default_args`
    default_owner = Airflow
    default_cpus = 1
    default_ram = 512
    default_disk = 512
    default_gpus = 0


    [webserver]
    # The base url of your website as airflow cannot guess what domain or
    # cname you are using. This is used in automated emails that
    # airflow sends to point links to the right web server
    base_url = http://localhost:8888

    # The ip specified when starting the web server
    web_server_host = 127.0.0.1

    # The port on which to run the web server
    web_server_port = 8888

    # Paths to the SSL certificate and key for the web server. When both are
    # provided SSL will be enabled. This does not change the web server port.
    web_server_ssl_cert =
    web_server_ssl_key =

    # Number of seconds the gunicorn webserver waits before timing out on a worker
    web_server_worker_timeout = 120

    # Number of workers to refresh at a time. When set to 0, worker refresh is
    # disabled. When nonzero, airflow periodically refreshes webserver workers by
    # bringing up new ones and killing old ones.
    worker_refresh_batch_size = 1

    # Number of seconds to wait before refreshing a batch of workers.
    worker_refresh_interval = 30

    # Secret key used to run your flask app
    secret_key = temporary_key

    # Number of workers to run the Gunicorn web server
    workers = 4

    # The worker class gunicorn should use. Choices include
    # sync (default), eventlet, gevent
    worker_class = sync

    # Log files for the gunicorn webserver. '-' means log to stderr.
    access_logfile = -
    error_logfile = -

    # Expose the configuration file in the web server
    expose_config = False

    # Set to true to turn on authentication:
    # http://pythonhosted.org/airflow/security.html#web-authentication
    authenticate = False

    # Filter the list of dags by owner name (requires authentication to be enabled)
    filter_by_owner = False

    # Filtering mode. Choices include user (default) and ldapgroup.
    # Ldap group filtering requires using the ldap backend
    #
    # Note that the ldap server needs the "memberOf" overlay to be set up
    # in order to user the ldapgroup mode.
    owner_mode = user

    # Default DAG orientation. Valid values are:
    # LR (Left->Right), TB (Top->Bottom), RL (Right->Left), BT (Bottom->Top)
    dag_orientation = LR

    # Puts the webserver in demonstration mode; blurs the names of Operators for
    # privacy.
    demo_mode = False

    # The amount of time (in secs) webserver will wait for initial handshake
    # while fetching logs from other worker machine
    log_fetch_timeout_sec = 5

    # By default, the webserver shows paused DAGs. Flip this to hide paused
    # DAGs by default
    hide_paused_dags_by_default = False

    [email]
    email_backend = airflow.utils.email.send_email_smtp


    [smtp]
    # If you want airflow to send emails on retries, failure, and you want to use
    # the airflow.utils.email.send_email_smtp function, you have to configure an
    # smtp server here
    smtp_host = localhost
    smtp_starttls = True
    smtp_ssl = False
    # Uncomment and set the user/pass settings if you want to use SMTP AUTH
    # smtp_user = airflow
    # smtp_password = airflow
    smtp_port = 25
    smtp_mail_from = [email protected]


    [celery]
    # This section only applies if you are using the CeleryExecutor in
    # [core] section above

    # The app name that will be used by celery
    celery_app_name = airflow.executors.celery_executor

    # The concurrency that will be used when starting workers with the
    # "airflow worker" command. This defines the number of task instances that
    # a worker will take, so size up your workers based on the resources on
    # your worker box and the nature of your tasks
    celeryd_concurrency = 16

    # When you start an airflow worker, airflow starts a tiny web server
    # subprocess to serve the workers local log files to the airflow main
    # web server, who then builds pages and sends them to users. This defines
    # the port on which the logs are served. It needs to be unused, and open
    # visible from the main web server to connect into the workers.
    worker_log_server_port = 8793

    # The Celery broker URL. Celery supports RabbitMQ, Redis and experimentally
    # a sqlalchemy database. Refer to the Celery documentation for more
    # information.
    broker_url = sqla+mysql://airflow:airflow@localhost:3306/airflow

    # Another key Celery setting
    celery_result_backend = db+mysql://airflow:airflow@localhost:3306/airflow

    # Celery Flower is a sweet UI for Celery. Airflow has a shortcut to start
    # it `airflow flower`. This defines the IP that Celery Flower runs on
    flower_host = 0.0.0.0

    # This defines the port that Celery Flower runs on
    flower_port = 5555

    # Default queue that tasks get assigned to and that worker listen on.
    default_queue = default


    [scheduler]
    # Task instances listen for external kill signal (when you clear tasks
    # from the CLI or the UI), this defines the frequency at which they should
    # listen (in seconds).
    job_heartbeat_sec = 5

    # The scheduler constantly tries to trigger new tasks (look at the
    # scheduler section in the docs for more information). This defines
    # how often the scheduler should run (in seconds).
    scheduler_heartbeat_sec = 5

    # after how much time should the scheduler terminate in seconds
    # -1 indicates to run continuously (see also num_runs)
    run_duration = -1

    # after how much time a new DAGs should be picked up from the filesystem
    min_file_process_interval = 0

    dag_dir_list_interval = 300

    # How often should stats be printed to the logs
    print_stats_interval = 30

    child_process_log_directory = /home/sherlock/workspace/airflow/logs/scheduler

    # Local task jobs periodically heartbeat to the DB. If the job has
    # not heartbeat in this many seconds, the scheduler will mark the
    # associated task instance as failed and will re-schedule the task.
    scheduler_zombie_task_threshold = 300

    # Turn off scheduler catchup by setting this to False.
    # Default behavior is unchanged and
    # Command Line Backfills still work, but the scheduler
    # will not do scheduler catchup if this is False,
    # however it can be set on a per DAG basis in the
    # DAG definition (catchup)
    catchup_by_default = True

    # Statsd (https://github.com/etsy/statsd) integration settings
    statsd_on = False
    statsd_host = localhost
    statsd_port = 8125
    statsd_prefix = airflow

    # The scheduler can run multiple threads in parallel to schedule dags.
    # This defines how many threads will run. However airflow will never
    # use more threads than the amount of cpu cores available.
    max_threads = 2

    authenticate = False


    [mesos]
    # Mesos master address which MesosExecutor will connect to.
    master = localhost:5050

    # The framework name which Airflow scheduler will register itself as on mesos
    framework_name = Airflow

    # Number of cpu cores required for running one task instance using
    # 'airflow run <dag_id> <task_id> <execution_date> --local -p <pickle_id>'
    # command on a mesos slave
    task_cpu = 1

    # Memory in MB required for running one task instance using
    # 'airflow run <dag_id> <task_id> <execution_date> --local -p <pickle_id>'
    # command on a mesos slave
    task_memory = 256

    # Enable framework checkpointing for mesos
    # See http://mesos.apache.org/documentation/latest/slave-recovery/
    checkpoint = False

    # Failover timeout in milliseconds.
    # When checkpointing is enabled and this option is set, Mesos waits
    # until the configured timeout for
    # the MesosExecutor framework to re-register after a failover. Mesos
    # shuts down running tasks if the
    # MesosExecutor framework fails to re-register within this timeframe.
    # failover_timeout = 604800

    # Enable framework authentication for mesos
    # See http://mesos.apache.org/documentation/latest/configuration/
    authenticate = False

    # Mesos credentials, if authentication is enabled
    # default_principal = admin
    # default_secret = admin


    [kerberos]
    ccache = /tmp/airflow_krb5_ccache
    # gets augmented with fqdn
    principal = airflow
    reinit_frequency = 3600
    kinit_path = kinit
    keytab = airflow.keytab


    [github_enterprise]
    api_rev = v3


    [admin]
    # UI to hide sensitive variable fields when set to True
    hide_sensitive_variable_fields = True
    104 changes: 104 additions & 0 deletions fetch_stock.py
    View file
    Original file line number Diff line number Diff line change
    @@ -0,0 +1,104 @@
    import json
    import csv
    import logging
    from pathlib import Path
    import argparse
    from collections import namedtuple

    import requests

    logging.basicConfig(level=logging.INFO)

    QUANDL_URL = 'https://www.quandl.com/api/v3/datasets/WIKI/{}.json'
    DATASTORE = Path('/project/data')

    Configuration = namedtuple(
    'Configuration',
    ['stock', 'start_date', 'end_date']
    )

    def fetch_stock_info(stock, start_date, end_date):
    response = requests.get(
    QUANDL_URL.format(stock),
    params={'start_date': start_date, 'end_date': end_date}
    )
    response.raise_for_status()
    json = response.json()
    data = json['dataset']['data']
    stock_price = {}
    for entry in data:
    date = entry[0]
    closing = entry[4]
    stock_price[date] = float(closing)
    return stock_price


    def read_current_data(store):
    stock_prices = {}
    try:
    with open(store) as f:
    reader = csv.reader(f)
    next(reader) # skip header
    for row in reader:
    date, price = row
    stock_prices[date] = price
    except FileNotFoundError:
    return {}
    return stock_prices


    def write_data(store, stock_prices):
    sorted_data = sorted(stock_prices.items(), key=lambda it: it[0])
    with open(store, 'w') as f:
    writer = csv.writer(f)
    writer.writerow(['date', 'price'])
    writer.writerows(sorted_data)


    def store_update(store, new_prices):
    stock_prices = read_current_data(store)
    stock_prices.update(new_prices)
    write_data(store, stock_prices)


    def fetch_and_update_stocks(stock, start_date, end_date):
    logging.info(
    f'Fetching stocks for symbol {stock} '
    f'between {start_date} and {end_date}'
    )
    prices = fetch_stock_info(stock, start_date, end_date)
    logging.info(f'Fetched {len(prices)} stock prices.')
    store = DATASTORE / f'{stock}.csv'
    store_update(store, prices)
    logging.info(f'Written information back to data store {store}.')


    def parse_command_line():
    parser = argparse.ArgumentParser(
    description='Fetch and store stock information')
    parser.add_argument('symbol')
    parser.add_argument(
    'start_date',
    help='start date, in format 2018-01-22'
    )
    parser.add_argument(
    'end_date',
    help='end date, in format 2018-01-29'
    )
    args = parser.parse_args()
    return Configuration(
    args.symbol.upper(),
    args.start_date,
    args.end_date
    )


    if __name__ == '__main__':
    configuration = parse_command_line()
    logging.info(f'Using configuration {configuration}.')
    fetch_and_update_stocks(
    configuration.stock,
    configuration.start_date,
    configuration.end_date
    )

    80 changes: 80 additions & 0 deletions publish_report.py
    View file
    Original file line number Diff line number Diff line change
    @@ -0,0 +1,80 @@
    import os

    import requests
    import sherlockml.auth

    import nbformat
    from nbconvert.preprocessors import ExecutePreprocessor


    HUDSON_URL = 'https://hudson.platform.asidata.science'
    TAVERN_URL = 'https://tavern.platform.asidata.science'

    REPORT_ID = 'b0b5e1c5-d7c9-4ae5-8b02-ce0f4a7ddaee'
    NOTEBOOK_PATH = '/project/stock-price/plot-stock-price.ipynb'


    class InputError(Exception):
    pass


    def run(report_id, notebook, run_notebook=True):
    """Publish a report in the current project.
    Parameters
    ----------
    report_id: str
    SherlockML report ID.
    notebook: str
    Full path to notebook on SherlockML.
    run_notebook: bool, optional
    Run notebook before publishing (default: True).
    """

    # Run the notebook
    if run_notebook:
    with open(notebook) as f:
    nb = nbformat.read(f, as_version=4)

    ep = ExecutePreprocessor(kernel_name='Python3')
    ep.preprocess(nb, {'metadata': {'path': os.path.dirname(notebook)}})

    with open(notebook, 'wt') as f:
    nbformat.write(nb, f)

    # Make sure notebook path starts with /project
    split_notebook_path = notebook.split('/')
    if split_notebook_path[0] == '' and split_notebook_path[1] == 'project':
    pass
    else:
    raise InputError('notebook path must start with /project')

    # Get rid of the project bit
    reconstructed_notebook_path = []
    for i, item in enumerate(split_notebook_path):
    if i < 2:
    continue
    reconstructed_notebook_path.append(item)
    reconstructed_notebook_path = '/'.join(reconstructed_notebook_path)

    # Authenticate and publish
    auth_headers = sherlockml.auth.auth_headers()
    user_id_response = requests.get(
    HUDSON_URL + '/authenticate',
    headers=auth_headers
    )
    user_id_response.raise_for_status()
    user_id = user_id_response.json()['account']['userId']
    body = {
    'notebook_path': reconstructed_notebook_path,
    'author_id': user_id,
    'draft': False
    }
    create_version_response = requests.post(
    TAVERN_URL + '/report/{}/version'.format(report_id),
    json=body,
    headers=auth_headers)
    create_version_response.raise_for_status()


    run(REPORT_ID, NOTEBOOK_PATH)
    35 changes: 35 additions & 0 deletions stock_price_dag.py
    View file
    Original file line number Diff line number Diff line change
    @@ -0,0 +1,35 @@
    from airflow import DAG
    from airflow.operators.bash_operator import BashOperator
    from datetime import datetime, timedelta

    default_args = {
    'owner': 'airflow',
    'depends_on_past': False,
    'start_date': datetime(2017, 1, 1),
    'email_on_failure': False,
    'email_on_retry': False,
    'retries': 1,
    'retry_delay': timedelta(minutes=5)
    }

    dag = DAG(
    'stock_price',
    default_args=default_args
    )


    fetch_command = '/opt/anaconda/envs/Python3/bin/python /project/stock-price/fetch_stock.py GOOG {{ ds }} {{ ds }}'

    fetch_operator = BashOperator(
    task_id='fetch_price',
    bash_command=fetch_command,
    dag = dag
    )

    publish_operator = BashOperator(
    task_id='publish',
    bash_command='/opt/anaconda/envs/Python3/bin/python /project/stock-price/publish.py',
    dag=dag
    )

    publish_operator.set_upstream(fetch_operator)
  23. pbugnion created this gist Jun 5, 2018.
    1 change: 1 addition & 0 deletions 01_readme.md
    View file
    Original file line number Diff line number Diff line change
    @@ -0,0 +1 @@
    # Airflow on SherlockML