Update readme, add examples

README.md

@@ -10,6 +10,10 @@ We are committed to supporting a baseline set of functionality on all x86 CPUs,
 
 ## Components
 
+### Config
+
+`Config` controls library-wide properties and must be passed to `Executor` and `DataMgr`. Default config objects should suffice for most installations. Instantiate a config first as part of library setup.
+
 ### Storage
 
 `ArrowStorage` is currently the default (and only available) HDK storage layer. `ArrowStorage` provides storage support for [Apache Arrow](https://github.com/apache/arrow) format data. The storage layer must be explicitly initialized:
@@ -72,6 +76,10 @@ The return from RelAlgExecutor is a ResultSet object which can be converted to A
      df = res.to_arrow().to_pandas()
 ```
 
+## Examples
+
+Standalone examples are available in the `examples` directory. Most examples run via Jupyter notebooks. 
+
 ## Build
 
 HDK includes components from external libraries as submodules. To clone the project, either use `git clone --recurse-submodules` to initially clone the repo, or clone without using flags and then run:

examples/basic_python_usage.ipynb

@@ -0,0 +1,98 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "fdef23a7",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Initialization\n",
+    "import pyhdk \n",
+    "\n",
+    "# Uses DBID 1\n",
+    "pyhdk.initLogger()\n",
+    "config = pyhdk.buildConfig()\n",
+    "storage = pyhdk.storage.ArrowStorage(1)\n",
+    "data_mgr = pyhdk.storage.DataMgr(config)\n",
+    "data_mgr.registerDataProvider(storage)\n",
+    "\n",
+    "calcite = pyhdk.sql.Calcite(storage, config)\n",
+    "executor = pyhdk.Executor(data_mgr, config)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "a5af996c",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Helper Functions\n",
+    "def get_rel_alg(sql):\n",
+    "    return calcite.process(sql)\n",
+    "\n",
+    "def run_query(sql):\n",
+    "    ra = get_rel_alg(sql)\n",
+    "    # One RelAlgExecutor per query\n",
+    "    rel_alg_executor = pyhdk.sql.RelAlgExecutor(executor, storage, data_mgr, ra)\n",
+    "    return rel_alg_executor.execute()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "3b771266",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "## Examples \n",
+    "\n",
+    "# Load some data\n",
+    "import pandas\n",
+    "import pyarrow as pa\n",
+    "\n",
+    "tbl = pa.Table.from_pandas(pandas.DataFrame({\"a\": [1, 2, 3], \"b\": [10, 20, 30]}))\n",
+    "opt = pyhdk.storage.TableOptions(2)\n",
+    "storage.importArrowTable(tbl, \"test\", opt)\n",
+    "\n",
+    "# Basic query\n",
+    "print(run_query(\"SELECT * FROM test;\").to_arrow().to_pandas())\n",
+    "\n",
+    "print(run_query(\"SELECT a, count(*), sum(b) FROM test GROUP BY a;\").to_arrow().to_pandas())\n",
+    "\n",
+    "# Cleanup\n",
+    "storage.dropTable(\"test\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "24e70d3f",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.13"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

scripts/conda/hdk-conda-dev-env.yml

@@ -0,0 +1 @@
+../../omniscidb/scripts/mapd-deps-conda-dev-env.yml