eval query support and a few fixes

api/py/ai/chronon/repo/interactive.md

@@ -22,29 +22,37 @@ Once we bulk it out - we can use this to power
 2. in your `~/.bashrc` or `~/.zshrc` add `SPARK_HOME` & `CHRONON_ROOT` env vars
 ```
 export SPARK_HOME=/Users/nikhilsimha/spark-3.5.4-bin-hadoop3                      # choose dir where you unpacked spark
-export CHRONON_ROOT=/Users/nikhilsimha/repos/chronon/api/py/test/sample   # choose your conf dir 
+export CHRONON_ROOT=/Users/nikhilsimha/repos/etsy/zipline   # choose your conf dir 
 ```
 
-### Running gateway
+3. Copy the chronon jar from your blob store and set it to `CHRONON_SPARK_JAR`.
+```
+# for EXAMPLE, in your ~/.zshrc
+export CHRONON_SPARK_JAR=/Users/nikhilsimha/Downloads/chronon_spark_assembly.jar
+```
 
-1.  Load parquet files into `$CHRONON_ROOT/local_warehouse/<namespace>/<table>`
 
+### Running gateway
+
+1. Load parquet files into `$CHRONON_ROOT/local_warehouse/<namespace>/<table>`
 
 2. Start gateway service - from chronon root
 ```
-> sbt spark/assembly && ./scripts/interactive/gateway.sh
+> ./scripts/interactive/gateway.sh
 ```
 
 ### Using interactively
 
 1. You can interact with your conf objects in vscode in a notebook like so..
 
 ```py
+
+# import source, I do this by running cell from the py file actually with "#%%" in vscode.
 from group_bys.etsy_search.visit_beacon import source
+# import runner
 from ai.chronon.repo.interactive import eval
 
-df = eval(source)
-df.show()
+eval(source)
 ```
 
 ```

api/py/ai/chronon/repo/interactive.py

@@ -4,7 +4,7 @@
 import ai.chronon.api.ttypes as thrift
 
 
-def eval(obj):
+def eval(obj, limit=5):
     """
     utility function to run conf's in an interactive environment.
 
@@ -22,7 +22,7 @@ def eval(obj):
     evaluator = gateway.entry_point
 
     if isinstance(obj, str):
-        evaluator.evalQuery(obj)  # TODO
+        return _to_df(evaluator.evalQuery(obj, limit))
 
     func = None
 
@@ -41,9 +41,12 @@ def eval(obj):
     elif isinstance(obj, thrift.Model):
         func = evaluator.evalModel  # TODO
 
+    else:
+        raise Exception(f"Unsupported object type for: {obj}")
+
     thrift_str = ser.thrift_simple_json(obj)
 
-    eval_result = func(thrift_str, 5)
+    eval_result = func(thrift_str, limit)
 
     return _to_df(eval_result)
 
@@ -57,4 +60,7 @@ def _to_df(eval_result):
         raise Exception(error)
 
     df = eval_result.df()
-    return DataFrame(df, df.sqlContext())
+
+    py_df = DataFrame(df, df.sparkSession())
+
+    return py_df

spark/src/main/scala/ai/chronon/spark/interactive/Evaluator.scala

@@ -33,6 +33,20 @@ class Evaluator(localWarehouse: LocalWarehouse) {
     }
   }
 
+  def evalQuery(query: String, limit: Int = 5): EvaluationResult = {
+
+    try {
+
+      EvaluationResult(null, localWarehouse.runSql(query).limit(limit))
+
+    } catch {
+
+      case analysisException: AnalysisException =>
+        EvaluationResult(analysisException.getMessage, null)
+
+    }
+  }
+
   private def runSourceBundle(sourceSqlBundle: SourceSqlBundle): DataFrame = {
 
     val missingTables = sourceSqlBundle.tables -- localWarehouse.existingTables