free text search

Author: jamesfisher-geo

dockerfiles/Dockerfile.dev.os

@@ -4,11 +4,10 @@ FROM python:3.10-slim
 # update apt pkgs, and install build-essential for ciso8601
 RUN apt-get update && \
     apt-get -y upgrade && \
-    apt-get -y install build-essential && \
+    apt-get -y install build-essential git && \
     apt-get clean && \
     rm -rf /var/lib/apt/lists/*
 
-RUN apt-get -y install git
 # update certs used by Requests
 ENV CURL_CA_BUNDLE=/etc/ssl/certs/ca-certificates.crt
 

stac_fastapi/core/stac_fastapi/core/extensions/query.py

@@ -9,6 +9,7 @@
 from enum import auto
 from types import DynamicClassAttribute
 from typing import Any, Callable, Dict, Optional
+import re
 
 from pydantic import BaseModel, model_validator
 from stac_pydantic.utils import AutoValueEnum

stac_fastapi/sfeos_helpers/stac_fastapi/sfeos_helpers/database/query.py

@@ -4,12 +4,74 @@
 """
 
 from typing import Any, Dict, List, Optional
+from stac_fastapi.types.rfc3339 import rfc3339_str_to_datetime
+from stac_fastapi.core.datetime_utils import datetime_to_str
 
 from stac_fastapi.sfeos_helpers.mappings import Geometry
 
 ES_MAX_URL_LENGTH = 4096
 
 
+def is_numeric(val: str) -> bool:
+    try:
+        float(val)
+        return True
+    except ValueError:
+        return False
+
+
+def process_ftq(q: str):
+    """Process a date, numeric, or text free-text query"""
+    q = q.strip()
+    if not q:
+        return
+    try:
+        if d := rfc3339_str_to_datetime(q):
+            return datetime_to_str(d)
+    except ValueError:
+        pass
+    if is_numeric(q):
+        return q
+    else:
+        return f"({q}* OR {q.lower()}* OR {q.upper()}*)"
+
+
+def apply_free_text_filter_shared(
+    search: Any,
+    free_text_queries: Optional[List[str]],
+    fields: Optional[List[str]] = [],
+) -> Any:
+    """Create a free text query for Elasticsearch/OpenSearch.
+
+    Args:
+        search (Any): The search object to apply the query to.
+        free_text_queries (Optional[List[str]]): A list of text strings to search for in the properties.
+        fields: Optional[List[str]]: A list of fields to return
+
+    Returns:
+        Any: The search object with the free text query applied, or the original search
+            object if no free_text_queries were provided.
+
+    Notes:
+        This function creates a query_string query that searches for the specified text strings
+        in the entire document. The query strings are joined with OR operators.
+    """
+
+    if free_text_queries:
+        processed_queries = [
+            process_ftq(q.strip()) for q in free_text_queries if q.strip()
+        ]
+
+        if processed_queries:
+            free_text_query_string = " AND ".join(processed_queries)
+
+            search = search.query(
+                "query_string", query=free_text_query_string, fields=fields
+            )
+
+    return search
+
+
 def apply_free_text_filter_shared(
     search: Any, free_text_queries: Optional[List[str]]
 ) -> Any: