From 70258e5c1d150a04c90a922310d46afe3ee2dd08 Mon Sep 17 00:00:00 2001
From: Taha Yassine Kraiem <tahayk2@gmail.com>
Date: Fri, 28 Mar 2025 16:00:27 +0100
Subject: [PATCH] refactor(chalice): simplified supportedTypes for product
 analytics

---
 .../core/product_analytics/events.py          |  4 +-
 .../core/product_analytics/properties.py      |  6 +-
 api/chalicelib/utils/exp_ch_helper.py         | 86 ++++++++++++++++++-
 3 files changed, 91 insertions(+), 5 deletions(-)

diff --git a/api/chalicelib/core/product_analytics/events.py b/api/chalicelib/core/product_analytics/events.py
index affccec7a..e645618bd 100644
--- a/api/chalicelib/core/product_analytics/events.py
+++ b/api/chalicelib/core/product_analytics/events.py
@@ -27,7 +27,7 @@ def get_events(project_id: int, page: schemas.PaginatedSchema):
     for i, row in enumerate(rows):
         row["id"] = f"event_{i}"
         row["icon"] = None
-        row["possibleTypes"] = ["String"]
+        row["possibleTypes"] = ["string"]
         row.pop("total")
     return {"total": total, "list": helper.list_to_camel_case(rows)}
 
@@ -141,6 +141,6 @@ def get_lexicon(project_id: int, page: schemas.PaginatedSchema):
     for i, row in enumerate(rows):
         row["id"] = f"event_{i}"
         row["icon"] = None
-        row["possibleTypes"] = ["String"]
+        row["possibleTypes"] = ["string"]
         row.pop("total")
     return {"total": total, "list": helper.list_to_camel_case(rows)}
diff --git a/api/chalicelib/core/product_analytics/properties.py b/api/chalicelib/core/product_analytics/properties.py
index e88f51de1..dcdfc954a 100644
--- a/api/chalicelib/core/product_analytics/properties.py
+++ b/api/chalicelib/core/product_analytics/properties.py
@@ -1,4 +1,4 @@
-from chalicelib.utils import helper
+from chalicelib.utils import helper, exp_ch_helper
 from chalicelib.utils.ch_client import ClickHouseClient
 import schemas
 
@@ -22,11 +22,13 @@ def get_all_properties(project_id: int, page: schemas.PaginatedSchema):
         if len(properties) == 0:
             return {"total": 0, "list": []}
         total = properties[0]["total"]
+        properties = helper.list_to_camel_case(properties)
         for i, p in enumerate(properties):
             p["id"] = f"prop_{i}"
             p["icon"] = None
+            p["possibleTypes"] = exp_ch_helper.simplify_clickhouse_types(p["possibleTypes"])
             p.pop("total")
-        return {"total": total, "list": helper.list_to_camel_case(properties)}
+        return {"total": total, "list": properties}
 
 
 def get_event_properties(project_id: int, event_name):
diff --git a/api/chalicelib/utils/exp_ch_helper.py b/api/chalicelib/utils/exp_ch_helper.py
index 15c4986d8..3b2ff7f2d 100644
--- a/api/chalicelib/utils/exp_ch_helper.py
+++ b/api/chalicelib/utils/exp_ch_helper.py
@@ -1,7 +1,8 @@
+import logging
+import re
 from typing import Union
 
 import schemas
-import logging
 
 logger = logging.getLogger(__name__)
 
@@ -66,3 +67,86 @@ def get_event_type(event_type: Union[schemas.EventType, schemas.PerformanceEvent
     if event_type not in defs:
         raise Exception(f"unsupported EventType:{event_type}")
     return defs.get(event_type)
+
+
+# AI generated
+def simplify_clickhouse_type(ch_type: str) -> str:
+    """
+    Simplify a ClickHouse data type name to a broader category like:
+    int, float, decimal, datetime, string, uuid, enum, array, tuple, map, nested, etc.
+    """
+
+    # 1) Strip out common wrappers like Nullable(...) or LowCardinality(...)
+    #    Possibly multiple wrappers: e.g. "LowCardinality(Nullable(Int32))"
+    pattern_wrappers = re.compile(r'(Nullable|LowCardinality)\((.*)\)')
+    while True:
+        match = pattern_wrappers.match(ch_type)
+        if match:
+            ch_type = match.group(2)
+        else:
+            break
+
+    # 2) Normalize (lowercase) for easier checks
+    normalized_type = ch_type.lower()
+
+    # 3) Use pattern matching or direct checks for known categories
+    #    (You can adapt this as you see fit for your environment.)
+
+    # Integers: Int8, Int16, Int32, Int64, Int128, Int256, UInt8, UInt16, ...
+    if re.match(r'^(u?int)(8|16|32|64|128|256)$', normalized_type):
+        return "int"
+
+    # Floats: Float32, Float64
+    if re.match(r'^float(32|64)$', normalized_type):
+        return "float"
+
+    # Decimal: Decimal(P, S)
+    if normalized_type.startswith("decimal"):
+        return "decimal"
+
+    # Date/DateTime
+    if normalized_type.startswith("date"):
+        return "datetime"
+    if normalized_type.startswith("datetime"):
+        return "datetime"
+
+    # Strings: String, FixedString(N)
+    if normalized_type.startswith("string"):
+        return "string"
+    if normalized_type.startswith("fixedstring"):
+        return "string"
+
+    # UUID
+    if normalized_type.startswith("uuid"):
+        return "uuid"
+
+    # Enums: Enum8(...) or Enum16(...)
+    if normalized_type.startswith("enum8") or normalized_type.startswith("enum16"):
+        return "enum"
+
+    # Arrays: Array(T)
+    if normalized_type.startswith("array"):
+        return "array"
+
+    # Tuples: Tuple(T1, T2, ...)
+    if normalized_type.startswith("tuple"):
+        return "tuple"
+
+    # Map(K, V)
+    if normalized_type.startswith("map"):
+        return "map"
+
+    # Nested(...)
+    if normalized_type.startswith("nested"):
+        return "nested"
+
+    # If we didn't match above, just return the original type in lowercase
+    return normalized_type
+
+
+def simplify_clickhouse_types(ch_types: list[str]) -> list[str]:
+    """
+    Takes a list of ClickHouse types and returns a list of simplified types
+    by calling `simplify_clickhouse_type` on each.
+    """
+    return list(set([simplify_clickhouse_type(t) for t in ch_types]))