test: cover search_with_snippets/5 and search_prefix/4

lib/muninn/searcher.ex

@@ -114,6 +114,10 @@ defmodule Muninn.Searcher do
     * `{:ok, results}` - Search results with total_hits and hits
     * `{:error, reason}` - Search or parse failed
 
+  > Note: `total_hits` is the number of hits returned in this response — capped by
+  > `:limit` and equal to `length(results["hits"])`, not the total number of
+  > matching documents. Use `count/3` for the full match count regardless of limit.
+
   ## Examples
 
       # Search for "elixir" in title and content fields
@@ -184,6 +188,10 @@ defmodule Muninn.Searcher do
     * `{:ok, results}` - Search results with snippets
     * `{:error, reason}` - Search or parse failed
 
+  > Note: `total_hits` is the number of hits returned in this response — capped by
+  > `:limit` and equal to `length(results["hits"])`, not the total number of
+  > matching documents. Use `count/3` for the full match count regardless of limit.
+
   Result format includes an additional `"snippets"` map with HTML-highlighted snippets:
 
       %{
@@ -275,6 +283,10 @@ defmodule Muninn.Searcher do
     * `{:ok, results}` - Search results with total_hits and hits
     * `{:error, reason}` - Search failed
 
+  > Note: `total_hits` is the number of hits returned in this response — capped by
+  > `:limit` and equal to `length(results["hits"])`, not the total number of
+  > matching documents. Use `count/3` for the full match count regardless of limit.
+
   ## Examples
 
       # Search for titles starting with "eli"

test/muninn/prefix_test.exs

@@ -0,0 +1,76 @@
+defmodule Muninn.PrefixTest do
+  use ExUnit.Case, async: true
+
+  alias Muninn.{Index, IndexWriter, IndexReader, Searcher, Schema}
+
+  setup do
+    test_path = "/tmp/muninn_prefix_#{:erlang.unique_integer([:positive])}"
+    on_exit(fn -> Muninn.TestHelpers.safe_rm_rf(test_path) end)
+    {:ok, test_path: test_path}
+  end
+
+  defp searcher_for(test_path, docs) do
+    schema = Schema.new() |> Schema.add_text_field("title", stored: true, indexed: true)
+    {:ok, index} = Index.create(test_path, schema)
+    Enum.each(docs, &IndexWriter.add_document(index, &1))
+    IndexWriter.commit(index)
+    {:ok, reader} = IndexReader.new(index)
+    {:ok, searcher} = Searcher.new(reader)
+    searcher
+  end
+
+  describe "search_prefix/4" do
+    test "matches documents whose term starts with the prefix", %{test_path: test_path} do
+      searcher =
+        searcher_for(test_path, [
+          %{"title" => "Phoenix Framework"},
+          %{"title" => "Photography Tips"},
+          %{"title" => "Elixir Guide"}
+        ])
+
+      {:ok, results} = Searcher.search_prefix(searcher, "title", "pho", limit: 10)
+
+      assert results["total_hits"] == 2
+      titles = Enum.map(results["hits"], & &1["doc"]["title"])
+      assert "Phoenix Framework" in titles
+      assert "Photography Tips" in titles
+      refute "Elixir Guide" in titles
+    end
+
+    test "narrows results as the prefix grows (typeahead)", %{test_path: test_path} do
+      searcher =
+        searcher_for(test_path, [
+          %{"title" => "program"},
+          %{"title" => "programming"},
+          %{"title" => "progress"},
+          %{"title" => "project"}
+        ])
+
+      {:ok, r1} = Searcher.search_prefix(searcher, "title", "pro", limit: 10)
+      {:ok, r2} = Searcher.search_prefix(searcher, "title", "progr", limit: 10)
+      {:ok, r3} = Searcher.search_prefix(searcher, "title", "program", limit: 10)
+
+      assert r1["total_hits"] == 4
+      assert r2["total_hits"] == 3
+      assert r3["total_hits"] == 2
+    end
+
+    test "respects the limit option", %{test_path: test_path} do
+      docs = for i <- 1..10, do: %{"title" => "testitem#{i}"}
+      searcher = searcher_for(test_path, docs)
+
+      {:ok, results} = Searcher.search_prefix(searcher, "title", "testitem", limit: 5)
+
+      assert length(results["hits"]) <= 5
+    end
+
+    test "returns no hits for a non-matching prefix", %{test_path: test_path} do
+      searcher = searcher_for(test_path, [%{"title" => "Elixir"}, %{"title" => "Phoenix"}])
+
+      {:ok, results} = Searcher.search_prefix(searcher, "title", "zzz", limit: 10)
+
+      assert results["total_hits"] == 0
+      assert results["hits"] == []
+    end
+  end
+end

test/muninn/snippets_test.exs

@@ -0,0 +1,115 @@
+defmodule Muninn.SnippetsTest do
+  use ExUnit.Case, async: true
+
+  alias Muninn.{Index, IndexWriter, IndexReader, Searcher, Schema}
+
+  setup do
+    test_path = "/tmp/muninn_snippets_#{:erlang.unique_integer([:positive])}"
+    on_exit(fn -> Muninn.TestHelpers.safe_rm_rf(test_path) end)
+    {:ok, test_path: test_path}
+  end
+
+  defp searcher_for(test_path, schema, docs) do
+    {:ok, index} = Index.create(test_path, schema)
+    Enum.each(docs, &IndexWriter.add_document(index, &1))
+    IndexWriter.commit(index)
+    {:ok, reader} = IndexReader.new(index)
+    {:ok, searcher} = Searcher.new(reader)
+    searcher
+  end
+
+  describe "search_with_snippets/5" do
+    test "returns a snippet map for the requested field", %{test_path: test_path} do
+      schema =
+        Schema.new()
+        |> Schema.add_text_field("title", stored: true, indexed: true)
+        |> Schema.add_text_field("content", stored: true, indexed: true)
+
+      searcher =
+        searcher_for(test_path, schema, [
+          %{"title" => "Elixir Guide", "content" => "learn elixir programming with this guide"}
+        ])
+
+      {:ok, results} =
+        Searcher.search_with_snippets(searcher, "elixir", ["title", "content"], ["content"])
+
+      assert results["total_hits"] == 1
+      hit = List.first(results["hits"])
+      assert is_map(hit["snippets"])
+      assert Map.has_key?(hit["snippets"], "content")
+      assert is_binary(hit["snippets"]["content"])
+    end
+
+    test "highlights the matched term with <b> tags", %{test_path: test_path} do
+      schema = Schema.new() |> Schema.add_text_field("content", stored: true, indexed: true)
+
+      searcher =
+        searcher_for(test_path, schema, [%{"content" => "learn elixir programming today"}])
+
+      {:ok, results} =
+        Searcher.search_with_snippets(searcher, "elixir", ["content"], ["content"])
+
+      snippet = results["hits"] |> List.first() |> get_in(["snippets", "content"])
+      assert snippet =~ "<b>elixir</b>"
+    end
+
+    test "supports multiple snippet fields", %{test_path: test_path} do
+      schema =
+        Schema.new()
+        |> Schema.add_text_field("title", stored: true, indexed: true)
+        |> Schema.add_text_field("content", stored: true, indexed: true)
+
+      searcher =
+        searcher_for(test_path, schema, [
+          %{"title" => "elixir basics", "content" => "an elixir tutorial"}
+        ])
+
+      {:ok, results} =
+        Searcher.search_with_snippets(
+          searcher,
+          "elixir",
+          ["title", "content"],
+          ["title", "content"]
+        )
+
+      snippets = results["hits"] |> List.first() |> Map.get("snippets")
+      assert snippets["title"] =~ "<b>elixir</b>"
+      assert snippets["content"] =~ "<b>elixir</b>"
+    end
+
+    test "max_snippet_chars truncates long content", %{test_path: test_path} do
+      schema = Schema.new() |> Schema.add_text_field("content", stored: true, indexed: true)
+
+      long =
+        String.duplicate("padding words here ", 40) <>
+          "elixir " <> String.duplicate("more padding text ", 40)
+
+      searcher = searcher_for(test_path, schema, [%{"content" => long}])
+
+      {:ok, results} =
+        Searcher.search_with_snippets(searcher, "elixir", ["content"], ["content"],
+          max_snippet_chars: 40
+        )
+
+      snippet = results["hits"] |> List.first() |> get_in(["snippets", "content"])
+      stripped = String.replace(snippet, ~r{</?b>}, "")
+
+      assert snippet =~ "<b>elixir</b>"
+      assert String.length(stripped) < String.length(long)
+    end
+
+    test "respects the limit option", %{test_path: test_path} do
+      schema = Schema.new() |> Schema.add_text_field("content", stored: true, indexed: true)
+
+      docs = for i <- 1..5, do: %{"content" => "elixir document number #{i}"}
+      searcher = searcher_for(test_path, schema, docs)
+
+      {:ok, results} =
+        Searcher.search_with_snippets(searcher, "elixir", ["content"], ["content"], limit: 2)
+
+      # total_hits reflects the number of hits actually returned (capped by limit)
+      assert results["total_hits"] == 2
+      assert length(results["hits"]) == 2
+    end
+  end
+end