Add file filtering with glob patterns to bake_folder

README.md

@@ -30,11 +30,75 @@ end
 ```
 
 **Options:**
-- `bake_folder(path, dir = __DIR__, allow_empty: false, include_dotfiles: false, max_size: nil)` - Bake all files in a directory
+- `bake_folder(path, dir = __DIR__, allow_empty: false, include_dotfiles: false, include_patterns: nil, exclude_patterns: nil, max_size: nil)` - Bake all files in a directory
 - `include_dotfiles: true` - Include files/folders starting with `.` (e.g., `.gitignore`)
 - `allow_empty: false` - Raise error if folder is empty
+- `include_patterns: Array(String)` - Include only files matching glob patterns
+- `exclude_patterns: Array(String)` - Exclude files matching glob patterns
 - `max_size: Int64` - Maximum total compressed size in bytes (compilation fails if exceeded)
 
+### File Filtering
+
+Use glob patterns to selectively include or exclude files when baking directories.
+
+**Pattern Syntax:**
+- `*` - Matches any characters except path separator (e.g., `*.cr` matches `file.cr`)
+- `**` - Matches zero or more directory levels (e.g., `**/*.cr` matches `src/file.cr`, `src/models/user.cr`)
+- `?` - Matches single character except path separator (e.g., `file?.txt` matches `file1.txt`)
+
+**Include Patterns** (whitelist approach):
+
+```crystal
+class Assets
+  extend BakedFileSystem
+
+  # Only include Crystal source files
+  bake_folder "./src", include_patterns: ["**/*.cr"]
+
+  # Include multiple file types
+  bake_folder "./docs", include_patterns: ["**/*.md", "**/*.txt"]
+end
+```
+
+**Exclude Patterns** (blacklist approach):
+
+```crystal
+class Assets
+  extend BakedFileSystem
+
+  # Exclude test and spec files
+  bake_folder "./project", exclude_patterns: ["**/test/*", "**/spec/*"]
+
+  # Exclude build artifacts and temporary files
+  bake_folder "./app", exclude_patterns: ["**/build/*", "**/*.tmp", "**/*.log"]
+end
+```
+
+**Combined Filtering** (include first, then exclude):
+
+```crystal
+class Assets
+  extend BakedFileSystem
+
+  # Include only source files, but exclude test files
+  bake_folder "./src",
+    include_patterns: ["**/*.cr", "**/*.md"],
+    exclude_patterns: ["**/test/*", "**/*_spec.cr"]
+end
+```
+
+**Important Notes:**
+- Patterns are relative to the baked directory (not absolute paths)
+- Include patterns are applied first (OR logic - match any pattern)
+- Exclude patterns are then applied (OR logic - exclude if matches any pattern)
+- Empty results raise error unless `allow_empty: true`
+
+**Use Cases:**
+- Embed only production assets (exclude dev/test files)
+- Include specific file types (source code, documentation)
+- Exclude large or generated files (builds, logs, cache)
+- Filter by directory structure (exclude vendor, node_modules)
+
 ### Loading Files
 
 ```crystal
@@ -126,7 +190,7 @@ crystal build your_app.cr
 - **Keep total embedded size under 50 MB** for reasonable binary sizes
 - **Use runtime loading for very large assets** (> 10 MB per file)
 - **Review compilation statistics** to catch accidentally embedded files
-- **Use `.gitignore`-style patterns** (future feature) to exclude build artifacts
+- **Use file filtering patterns** to exclude build artifacts, tests, and development files
 - **Monitor benchmark results** to understand compile time and binary size impact
 
 ### Advanced

benchmarks/binary_size.cr

@@ -5,8 +5,8 @@ require "json"
 
 module BinarySizeBenchmark
   BASELINE_DIR = File.expand_path("baseline", __DIR__)
-  BAKED_DIR = File.expand_path("baked", __DIR__)
-  PUBLIC_DIR = File.expand_path("public", __DIR__)
+  BAKED_DIR    = File.expand_path("baked", __DIR__)
+  PUBLIC_DIR   = File.expand_path("public", __DIR__)
   RESULTS_FILE = File.expand_path("results/binary_size.json", __DIR__)
 
   struct BinaryInfo

benchmarks/compile_time.cr

@@ -5,9 +5,9 @@ require "file_utils"
 # Tests compilation time overhead of embedding assets
 
 module CompileTimeBenchmark
-  ITERATIONS = 5
+  ITERATIONS   = 5
   BASELINE_DIR = File.expand_path("baseline", __DIR__)
-  BAKED_DIR = File.expand_path("baked", __DIR__)
+  BAKED_DIR    = File.expand_path("baked", __DIR__)
   RESULTS_FILE = File.expand_path("results/compile_time.json", __DIR__)
 
   struct CompileResult

benchmarks/memory.cr

@@ -5,12 +5,12 @@ require "http/client"
 # Measures RSS (Resident Set Size) at various stages
 
 module MemoryBenchmark
-  BASELINE_DIR = File.expand_path("baseline", __DIR__)
-  BAKED_DIR = File.expand_path("baked", __DIR__)
-  RESULTS_FILE = File.expand_path("results/memory.json", __DIR__)
+  BASELINE_DIR  = File.expand_path("baseline", __DIR__)
+  BAKED_DIR     = File.expand_path("baked", __DIR__)
+  RESULTS_FILE  = File.expand_path("results/memory.json", __DIR__)
   BASELINE_PORT = 3000
-  BAKED_PORT = 3001
-  WARMUP_DELAY = 2 # seconds to wait for server startup
+  BAKED_PORT    = 3001
+  WARMUP_DELAY  =    2 # seconds to wait for server startup
 
   struct MemoryProfile
     include JSON::Serializable

benchmarks/performance.cr

@@ -5,14 +5,14 @@ require "http/client"
 # Measures request latency and throughput
 
 module PerformanceBenchmark
-  BASELINE_DIR = File.expand_path("baseline", __DIR__)
-  BAKED_DIR = File.expand_path("baked", __DIR__)
-  RESULTS_FILE = File.expand_path("results/performance.json", __DIR__)
-  BASELINE_PORT = 3000
-  BAKED_PORT = 3001
-  WARMUP_REQUESTS = 100
+  BASELINE_DIR       = File.expand_path("baseline", __DIR__)
+  BAKED_DIR          = File.expand_path("baked", __DIR__)
+  RESULTS_FILE       = File.expand_path("results/performance.json", __DIR__)
+  BASELINE_PORT      = 3000
+  BAKED_PORT         = 3001
+  WARMUP_REQUESTS    =  100
   BENCHMARK_REQUESTS = 1000
-  CONCURRENT_CLIENTS = 10
+  CONCURRENT_CLIENTS =   10
 
   struct LatencyStats
     include JSON::Serializable

examples/threshold_example.cr

@@ -18,7 +18,7 @@ module Assets
   # - medium.bin: 50KB → ~50KB compressed
   # - large.bin: 100KB → ~100KB compressed
   # Total: ~150KB compressed
-  bake_folder path: "./assets", max_size: 51_200  # 50 KB limit - will fail!
+  bake_folder path: "./assets", max_size: 51_200 # 50 KB limit - will fail!
 end
 
 # This code will never execute because compilation fails above

spec/baked_file_system_spec.cr

@@ -43,6 +43,37 @@ class EdgeCaseStorage
   bake_folder "./storage_edge_cases"
 end
 
+# Test storage with include patterns - only .cr files
+class FilteredStorageInclude
+  extend BakedFileSystem
+  bake_folder "./storage/filters", include_patterns: ["**/*.cr"]
+end
+
+# Test storage with exclude patterns - exclude test directory
+class FilteredStorageExclude
+  extend BakedFileSystem
+  bake_folder "./storage/filters", exclude_patterns: ["**/test/*"]
+end
+
+# Test storage with both include and exclude patterns
+class FilteredStorageCombined
+  extend BakedFileSystem
+  bake_folder "./storage/filters", include_patterns: ["**/*.cr", "**/*.md"], exclude_patterns: ["**/test/*"]
+end
+
+# Test storage with patterns that result in empty set (unless allow_empty)
+class FilteredStorageEmpty
+  extend BakedFileSystem
+  bake_folder "./storage/filters", include_patterns: ["**/*.txt"], allow_empty: true
+end
+
+# This should raise an error at compile time - patterns match nothing and allow_empty is false
+# Commented out because it would prevent compilation
+# class FilteredStorageEmptyError
+#   extend BakedFileSystem
+#   bake_folder "./storage/filters", include_patterns: ["**/*.txt"]
+# end
+
 def read_slice(path)
   File.open(path, "rb") do |io|
     Slice(UInt8).new(io.size).tap do |buf|
@@ -53,7 +84,7 @@ end
 
 describe BakedFileSystem do
   it "load only files without hidden one" do
-    Storage.files.size.should eq(4)
+    Storage.files.size.should eq(10) # lorem.txt, images/sidekiq.png, string_encoding/*, filters/*
     Storage.get?(".hidden/hidden_file.txt").should be_nil
   end
 
@@ -115,7 +146,7 @@ describe BakedFileSystem do
   end
 
   it "handles interpolation in content" do
-    String.new(Storage.get("string_encoding/interpolation.gz").to_slice).should eq "\#{foo} \{% macro %}\n"
+    String.new(Storage.get("string_encoding/interpolation.gz").to_slice).should eq "\#{foo} {% macro %}\n"
   end
 
   describe "rewind functionality" do
@@ -580,4 +611,63 @@ describe BakedFileSystem do
       file.not_nil!.closed?.should be_true
     end
   end
+
+  describe "file filtering" do
+    it "includes only files matching include patterns" do
+      # Should only have .cr files
+      FilteredStorageInclude.files.size.should eq(4) # src/main.cr, src/lib.cr, test/spec.cr, test/helper.cr
+      FilteredStorageInclude.get?("src/main.cr").should_not be_nil
+      FilteredStorageInclude.get?("src/lib.cr").should_not be_nil
+      FilteredStorageInclude.get?("test/spec.cr").should_not be_nil
+      FilteredStorageInclude.get?("test/helper.cr").should_not be_nil
+      FilteredStorageInclude.get?("docs/README.md").should be_nil
+      FilteredStorageInclude.get?("config.yml").should be_nil
+    end
+
+    it "excludes files matching exclude patterns" do
+      # Should have everything except test/* files
+      FilteredStorageExclude.files.size.should eq(4) # src/*, docs/*, config.yml
+      FilteredStorageExclude.get?("src/main.cr").should_not be_nil
+      FilteredStorageExclude.get?("src/lib.cr").should_not be_nil
+      FilteredStorageExclude.get?("docs/README.md").should_not be_nil
+      FilteredStorageExclude.get?("config.yml").should_not be_nil
+      FilteredStorageExclude.get?("test/spec.cr").should be_nil
+      FilteredStorageExclude.get?("test/helper.cr").should be_nil
+    end
+
+    it "applies both include and exclude patterns" do
+      # Include *.cr and *.md, exclude test/*
+      # Should have: src/*.cr and docs/*.md (not test/*.cr)
+      FilteredStorageCombined.files.size.should eq(3) # src/main.cr, src/lib.cr, docs/README.md
+      FilteredStorageCombined.get?("src/main.cr").should_not be_nil
+      FilteredStorageCombined.get?("src/lib.cr").should_not be_nil
+      FilteredStorageCombined.get?("docs/README.md").should_not be_nil
+      FilteredStorageCombined.get?("test/spec.cr").should be_nil
+      FilteredStorageCombined.get?("test/helper.cr").should be_nil
+      FilteredStorageCombined.get?("config.yml").should be_nil
+    end
+
+    it "handles empty result with allow_empty" do
+      # No .txt files in filters directory, but allow_empty is true
+      FilteredStorageEmpty.files.size.should eq(0)
+    end
+
+    it "filters are relative to baked directory" do
+      # Patterns should match from the baked folder root, not absolute paths
+      FilteredStorageInclude.get?("src/main.cr").should_not be_nil
+      # Not "/storage/filters/src/main.cr"
+    end
+
+    it "can read content from filtered files" do
+      file = FilteredStorageInclude.get("src/main.cr")
+      content = file.gets_to_end
+      content.should contain("Main file")
+    end
+
+    it "empty result behavior respects allow_empty flag" do
+      # FilteredStorageEmpty has allow_empty: true, so it should work with 0 files
+      FilteredStorageEmpty.files.size.should eq(0)
+      # Without allow_empty: true, it would raise at compile time (tested manually)
+    end
+  end
 end