feat(athena): add amazon athena backend

.github/workflows/ibis-backends-cloud.yml

@@ -46,6 +46,10 @@ jobs:
             title: Databricks
             extras:
               - --extra databricks
+          - name: athena
+            title: Amazon Athena
+            extras:
+              - --extra athena
         include:
           - python-version: "3.10"
             backend:
@@ -78,7 +82,7 @@ jobs:
     # rate limit while restricting GITHUB_TOKEN permissions elsewhere
     permissions:
       contents: "read"
-      # required for GCP workload identity federation
+      # required for workload identity federation
       id-token: "write"
 
     steps:
@@ -127,6 +131,7 @@ jobs:
         run: just download-data
 
       - uses: google-github-actions/auth@v2
+        if: matrix.backend.name == 'bigquery'
         with:
           project_id: "ibis-gbq"
           workload_identity_provider: "${{ vars.WIF_PROVIDER_NAME }}"
@@ -164,6 +169,13 @@ jobs:
           SNOWFLAKE_SCHEMA: ${{ secrets.SNOWFLAKE_SCHEMA }}
           SNOWFLAKE_WAREHOUSE: ${{ secrets.SNOWFLAKE_WAREHOUSE }}
 
+      - name: setup aws credentials
+        if: matrix.backend.name == 'athena'
+        uses: aws-actions/configure-aws-credentials@v4
+        with:
+          aws-region: us-east-2
+          role-to-assume: arn:aws:iam::070284473168:role/ibis-project-athena
+
       - name: enable snowpark testing
         if: matrix.backend.key == 'snowpark'
         run: echo "SNOWFLAKE_SNOWPARK=1" >> "$GITHUB_ENV"

docs/backends/athena.qmd

@@ -0,0 +1,117 @@
+# Amazon Athena
+
+[https://aws.amazon.com/athena](https://aws.amazon.com/athena/)
+
+## Install
+
+Install Ibis and dependencies for the Athena backend:
+
+::: {.panel-tabset}
+
+## `pip`
+
+Install with the `athena` extra:
+
+```{.bash}
+pip install 'ibis-framework[athena]'
+```
+
+And connect:
+
+```{.python}
+import ibis
+
+con = ibis.athena.connect(s3_staging_dir="s3://...")  # <1>
+```
+
+::: {.callout-note}
+## At a **minimum**, the `s3_staging_dir` argument must be provided.
+
+This argument tells the underlying driver library--`pyathena`--and ultimately
+Athena itself where to dump query results.
+:::
+
+1. Adjust other connection parameters as needed.
+
+## `conda`
+
+Install for Athena:
+
+```{.bash}
+conda install -c conda-forge ibis-athena
+```
+
+```{.python}
+import ibis
+
+con = ibis.athena.connect(s3_staging_dir="s3://...")  # <1>
+```
+
+::: {.callout-note}
+## At a **minimum**, the `s3_staging_dir` argument must be provided.
+
+This argument tells the underlying driver library--`pyathena`--and ultimately
+Athena itself where to dump query results.
+:::
+
+1. Adjust other connection parameters as needed.
+
+## `mamba`
+
+Install for Athena:
+
+```{.bash}
+mamba install -c conda-forge ibis-athena
+```
+
+```{.python}
+import ibis
+
+con = ibis.athena.connect(s3_staging_dir="s3://my-bucket/")  # <1>
+```
+
+::: {.callout-note}
+## At a **minimum**, the `s3_staging_dir` argument must be provided.
+
+This argument tells the underlying driver library--`pyathena`--and ultimately
+Athena itself where to dump query results.
+:::
+
+1. Adjust other connection parameters as needed.
+
+:::
+
+## Connect
+
+### `ibis.athena.connect`
+
+```python
+con = ibis.athena.connect(
+    s3_staging_dir="s3://my-bucket/",
+)
+```
+
+::: {.callout-note}
+## At a **minimum**, the `s3_staging_dir` argument must be provided.
+
+This argument tells the underlying driver
+library--[`pyathena`](https://laughingman7743.github.io/PyAthena/)--and
+ultimately Athena itself where to dump query results.
+:::
+
+### Connection Parameters
+
+```{python}
+#| echo: false
+#| output: asis
+from _utils import render_do_connect
+
+render_do_connect("athena")
+```
+
+```{python}
+#| echo: false
+BACKEND = "Athena"
+```
+
+{{< include ./_templates/api.qmd >}}

docs/backends/support/cloud_support_policy.qmd

@@ -1,16 +1,26 @@
 # Cloud backend support policy
 
-Ibis supports a number of proprietary cloud backends.
+Ibis supports and tests some of the major proprietary cloud backends:
 
-Snowflake and Databricks cost money to test in our continuous integration test suite.
+- Google BigQuery
+- Snowflake
+- Databricks Cloud
+- Amazon Athena
 
-If funding dries up, support for these backends will become best effort.
+**All of these services cost money to run in our continuous integration test suite.**
 
-If you're interested in ensuring continued support for a cloud backend, please
-consider sponsoring Ibis development.
+If funding for these services dries up, support for these backends will move to
+best effort.
 
-The cost at the time of writing (2024-10-31) is about **$5,000 USD per year**
-split evenly between the Snowflake and Databricks backend.
+If you or your company are interested in ensuring continued maintenance for one
+of these backends, please get in touch with us on
+[Zulip](https://ibis-project.zulipchat.com/), or [open a GitHub
+issue](https://github.com/ibis-project/ibis/issues/new) and consider sponsoring
+Ibis development.
 
-Google has very generously supported the entire cost of testing the BigQuery
-backend for a number of years.
+## Current cost estimates
+
+- Google BigQuery: free, donated very generously by Google
+- Snowflake: $2500 USD per year
+- Databricks Cloud: $2500 USD per year
+- Amazon Athena: ?

flake.nix

@@ -143,6 +143,7 @@
           inherit shellHook;
 
           PYSPARK_PYTHON = "${env}/bin/python";
+          AWS_PROFILE = "ibis-testing";
 
           # needed for mssql+pyodbc
           ODBCSYSINI = pkgs.writeTextDir "odbcinst.ini" ''