Release 4.4.46

Adds `.gitignore` rules to exclude benchmark outputs and local build artifacts. Initializes Serena AI configuration for the Go project with UTF-8 encoding and gitignore-based file exclusion. Refactors the FASTSEQ header parser to fix regex syntax, optimize alternation ordering, improve control flow readability, and introduce explicit float-to-int coercion that preserves fractional values.

.gitignore

@@ -21,6 +21,8 @@ xx
 
 .rhistory
 /.vscode
+
+/benchmarck
 /build
 /bugs
 autodoc

.serena/.gitignore

@@ -0,0 +1,2 @@
+/cache
+/project.local.yml

.serena/project.yml

@@ -0,0 +1,133 @@
+# the name by which the project can be referenced within Serena
+project_name: "obitools4"
+
+
+# list of languages for which language servers are started; choose from:
+#   al                  angular             ansible             bash                clojure
+#   cpp                 cpp_ccls            crystal             csharp              csharp_omnisharp
+#   dart                elixir              elm                 erlang              fortran
+#   fsharp              go                  groovy              haskell             haxe
+#   hlsl                html                java                json                julia
+#   kotlin              lean4               lua                 luau                markdown
+#   matlab              msl                 nix                 ocaml               pascal
+#   perl                php                 php_phpactor        powershell          python
+#   python_jedi         python_ty           r                   rego                ruby
+#   ruby_solargraph     rust                scala               scss                solidity
+#   svelte              swift               systemverilog       terraform           toml
+#   typescript          typescript_vts      vue                 yaml                zig
+#   (This list may be outdated. For the current list, see values of Language enum here:
+#   https://github.com/oraios/serena/blob/main/src/solidlsp/ls_config.py
+#   For some languages, there are alternative language servers, e.g. csharp_omnisharp, ruby_solargraph.)
+# Note:
+#   - For C, use cpp
+#   - For JavaScript, use typescript
+#   - For Angular projects, use angular (subsumes typescript+html; requires `npm install` in the project root)
+#   - For Svelte projects, use svelte (subsumes typescript/javascript for .svelte projects; requires npm)
+#   - For SCSS / Sass / plain CSS, use scss (some-sass-language-server handles all three)
+#   - For Free Pascal/Lazarus, use pascal
+# Special requirements:
+#   Some languages require additional setup/installations.
+#   See here for details: https://oraios.github.io/serena/01-about/020_programming-languages.html#language-servers
+# When using multiple languages, the first language server that supports a given file will be used for that file.
+# The first language is the default language and the respective language server will be used as a fallback.
+# Note that when using the JetBrains backend, language servers are not used and this list is correspondingly ignored.
+languages:
+- go
+
+# the encoding used by text files in the project
+# For a list of possible encodings, see https://docs.python.org/3.11/library/codecs.html#standard-encodings
+encoding: "utf-8"
+
+# line ending convention to use when writing source files.
+# Possible values: unset (use global setting), "lf", "crlf", or "native" (platform default)
+# This does not affect Serena's own files (e.g. memories and configuration files), which always use native line endings.
+line_ending:
+
+# The language backend to use for this project.
+# If not set, the global setting from serena_config.yml is used.
+# Valid values: LSP, JetBrains
+# Note: the backend is fixed at startup. If a project with a different backend
+# is activated post-init, an error will be returned.
+language_backend:
+
+# whether to use project's .gitignore files to ignore files
+ignore_all_files_in_gitignore: true
+
+# advanced configuration option allowing to configure language server-specific options.
+# Maps the language key to the options.
+# Have a look at the docstring of the constructors of the LS implementations within solidlsp (e.g., for C# or PHP) to see which options are available.
+# No documentation on options means no options are available.
+ls_specific_settings: {}
+
+# list of additional workspace folder paths for cross-package reference support (e.g. in monorepos).
+# Paths can be absolute or relative to the project root.
+# Each folder is registered as an LSP workspace folder, enabling language servers to discover
+# symbols and references across package boundaries.
+# Currently supported for: TypeScript.
+# Example:
+#   additional_workspace_folders:
+#     - ../sibling-package
+#     - ../shared-lib
+additional_workspace_folders: []
+
+# list of additional paths to ignore in this project.
+# Same syntax as gitignore, so you can use * and **.
+# Note: global ignored_paths from serena_config.yml are also applied additively.
+ignored_paths: []
+
+# whether the project is in read-only mode
+# If set to true, all editing tools will be disabled and attempts to use them will result in an error
+# Added on 2025-04-18
+read_only: false
+
+# list of tool names to exclude.
+# This extends the existing exclusions (e.g. from the global configuration)
+# Find the list of tools here: https://oraios.github.io/serena/01-about/035_tools.html
+excluded_tools: []
+
+# list of tools to include that would otherwise be disabled (particularly optional tools that are disabled by default).
+# This extends the existing inclusions (e.g. from the global configuration).
+# Find the list of tools here: https://oraios.github.io/serena/01-about/035_tools.html
+included_optional_tools: []
+
+# fixed set of tools to use as the base tool set (if non-empty), replacing Serena's default set of tools.
+# This cannot be combined with non-empty excluded_tools or included_optional_tools.
+# Find the list of tools here: https://oraios.github.io/serena/01-about/035_tools.html
+fixed_tools: []
+
+# list of mode names that are to be activated by default, overriding the setting in the global configuration.
+# The full set of modes to be activated is base_modes (from global config) + default_modes + added_modes.
+# If the setting is undefined/empty, the default_modes from the global configuration (serena_config.yml) apply.
+# Otherwise, this overrides the setting from the global configuration (serena_config.yml).
+# Therefore, you can set this to [] if you do not want the default modes defined in the global config to apply
+# for this project.
+# This setting can, in turn, be overridden by CLI parameters (--mode).
+# See https://oraios.github.io/serena/02-usage/050_configuration.html#modes
+default_modes:
+
+# list of mode names to be activated additionally for this project, e.g. ["query-projects"]
+# The full set of modes to be activated is base_modes (from global config) + default_modes + added_modes.
+# See https://oraios.github.io/serena/02-usage/050_configuration.html#modes
+added_modes:
+
+# initial prompt for the project. It will always be given to the LLM upon activating the project
+# (contrary to the memories, which are loaded on demand).
+initial_prompt: ""
+
+# time budget (seconds) per tool call for the retrieval of additional symbol information
+# such as docstrings or parameter information.
+# This overrides the corresponding setting in the global configuration; see the documentation there.
+# If null or missing, use the setting from the global configuration.
+symbol_info_budget:
+
+# list of regex patterns which, when matched, mark a memory entry as read‑only.
+# Extends the list from the global configuration, merging the two lists.
+read_only_memory_patterns: []
+
+# list of regex patterns for memories to completely ignore.
+# Matching memories will not appear in list_memories or activate_project output
+# and cannot be accessed via read_memory or write_memory.
+# To access ignored memory files, use the read_file tool on the raw file path.
+# Extends the list from the global configuration, merging the two lists.
+# Example: ["_archive/.*", "_episodes/.*"]
+ignored_memory_patterns: []

pkg/obiformats/fastseq_obi_header.go

@@ -17,7 +17,7 @@ import (
 )
 
 var __obi_header_value_string_pattern__ = regexp.MustCompile(`^'\s*([^']*'|"[^"]*")\s*;`)
-var __obi_header_value_numeric_pattern__ = regexp.MustCompile(`^\s*([+-]?\.\d+|[+-]?\d+(\.\d*)?([eE][+-]?\d+)?)\s*;`)
+var __obi_header_value_numeric_pattern__ = regexp.MustCompile(`^\s*[+-]?(\d+(\.\d*)?|\.\d+)([eE][+-]?\d+)?\s*;`)
 var __obi_header_map_int_key__ = regexp.MustCompile("([{,])([0-9]+):")
 
 func __match__dict__(text []byte) []int {
@@ -146,6 +146,65 @@ func __match__key__(text []byte) []int {
 	return []int{} // Not a key
 }
 
+func __match__array__(text []byte) []int {
+
+	state := 0
+	level := 0
+	start := 0
+	instring := byte(0)
+
+	for i, r := range text {
+		if state == 2 {
+			if r == ';' {
+				return []int{start, i + 1}
+			}
+			if r != ' ' && r != '\t' {
+				return []int{}
+			}
+		}
+
+		if state == 0 {
+			if r == '[' {
+				level++
+				state++
+				start = i
+				continue
+			}
+			if r != ' ' && r != '\t' {
+				return []int{}
+			}
+			continue
+		}
+
+		// state == 1: inside the array
+		if instring != 0 {
+			if r == instring {
+				instring = 0
+			}
+			continue
+		}
+
+		if r == '"' || r == '\'' {
+			instring = r
+			continue
+		}
+
+		if r == '[' || r == '{' {
+			level++
+			continue
+		}
+
+		if r == ']' || r == '}' {
+			level--
+			if level == 0 {
+				state++
+			}
+		}
+	}
+
+	return []int{}
+}
+
 func __match__general__(text []byte) []int {
 
 	for i, r := range text {
@@ -242,6 +301,21 @@ func ParseOBIFeatures(text string, annotations obiseq.Annotation) string {
 					stop = m[1] + 1
 				} else {
 
+					// array value
+					m = __match__array__(part)
+					if len(m) > 0 {
+						bvalue = bytes.TrimSpace(part[m[0]:(m[1] - 1)])
+						j := bytes.ReplaceAll(bvalue, []byte("'"), []byte(`"`))
+						j = __obi_header_map_int_key__.ReplaceAll(j, []byte(`$1"$2":`))
+						arr, err := _parse_json_array_interface(j)
+						if err != nil {
+							value = string(bvalue)
+						} else {
+							value = arr
+						}
+						stop = m[1] + 1
+					} else {
+
 						// Generic value
 
 						// m = __obi_header_value_general_pattern__.FindIndex(part)
@@ -264,6 +338,7 @@ func ParseOBIFeatures(text string, annotations obiseq.Annotation) string {
 							// no value
 							break
 						} // End of No value
+					} // End of not array
 				} // End of not dict
 			} // End of not string
 		} // End of not numeric
@@ -272,6 +347,8 @@ func ParseOBIFeatures(text string, annotations obiseq.Annotation) string {
 		case float64:
 			if vt == math.Floor(vt) {
 				annotations[key] = int(vt)
+			} else {
+				annotations[key] = vt
 			}
 		default:
 			annotations[key] = value
@@ -327,9 +404,8 @@ func WriteFastSeqOBIHeade(buffer *bytes.Buffer, sequence *obiseq.BioSequence) {
 					buffer.WriteString(fmt.Sprintf("%s=", key))
 					buffer.Write(tv)
 					buffer.WriteString("; ")
-				case map[string]int,
+				default:
-					map[string]string,
+					if obiutils.IsAMap(value) || obiutils.IsASlice(value) || obiutils.IsAnArray(value) {
-					map[string]interface{}:
 						tv, err := obiutils.JsonMarshal(t)
 						if err != nil {
 							log.Fatalf("Cannot convert %v value", value)
@@ -338,11 +414,12 @@ func WriteFastSeqOBIHeade(buffer *bytes.Buffer, sequence *obiseq.BioSequence) {
 						buffer.WriteString(fmt.Sprintf("%s=", key))
 						buffer.Write(tv)
 						buffer.WriteString("; ")
-				default:
+					} else {
 						buffer.WriteString(fmt.Sprintf("%s=%v; ", key, value))
 					}
 				}
 			}
+		}
 
 		if sequence.HasDefinition() {
 			buffer.WriteByte(' ')

pkg/obiformats/fastseq_write_fasta.go

@@ -90,6 +90,9 @@ func FormatFastaBatch(batch obiiter.BioSequenceBatch, formater FormatHeader, ski
 	log.Debugf("FormatFastaBatch: #%d : %d seqs", batch.Order(), batch.Len())
 
 	for _, seq := range batch.Slice() {
+		if len(seq.Id()) == 0 {
+			log.Fatalf("Sequence identifier is empty")
+		}
 		if seq.Len() > 0 {
 			// Write header directly into bs — no intermediate string
 			bs.WriteByte('>')

pkg/obiformats/fastseq_write_fastq.go

@@ -64,6 +64,9 @@ func FormatFastqBatch(batch obiiter.BioSequenceBatch,
 	first := true
 
 	for _, seq := range batch.Slice() {
+		if len(seq.Id()) == 0 {
+			log.Fatalf("Sequence identifier is empty")
+		}
 		if seq.Len() > 0 {
 			_formatFastq(&bs, seq, formater)
 

pkg/obikmer/entropy.go

@@ -4,22 +4,21 @@ import "math"
 
 // KmerEntropy computes the entropy of a single encoded k-mer.
 //
-// The algorithm mirrors the lowmask entropy calculation: it decodes the k-mer
+// The algorithm mirrors the Rust obiskbuilder entropy: it decodes the k-mer
 // to a DNA sequence, extracts all sub-words of each size from 1 to levelMax,
 // normalizes them by circular canonical form, counts their frequencies, and
-// computes Shannon entropy normalized by the maximum possible entropy.
+// computes Shannon entropy corrected for class sizes, normalized by the
+// maximum possible entropy over 4^ws raw bins.
 // The returned value is the minimum entropy across all word sizes.
 //
+// Correction for small sequences: the raw entropy H = log(N) - Σ f·log(f)/N
+// under-estimates the true complexity when many raw words collapse to the same
+// canonical form.  Adding Σ f·log(class_size)/N recovers the entropy of the
+// underlying uncollapsed distribution (assuming uniform mixing within each
+// equivalence class).
+//
 // A value close to 0 indicates very low complexity (e.g. "AAAA..."),
 // while a value close to 1 indicates high complexity.
-//
-// Parameters:
-//   - kmer: the encoded k-mer (2 bits per base)
-//   - k: the k-mer size
-//   - levelMax: maximum sub-word size for entropy (typically 6)
-//
-// Returns:
-//   - minimum normalized entropy across all word sizes 1..levelMax
 func KmerEntropy(kmer uint64, k int, levelMax int) float64 {
 	if k < 1 || levelMax < 1 {
 		return 1.0
@@ -35,7 +34,7 @@ func KmerEntropy(kmer uint64, k int, levelMax int) float64 {
 	var seqBuf [32]byte
 	seq := DecodeKmer(kmer, k, seqBuf[:])
 
-	// Pre-compute nLogN lookup (same as lowmask)
+	// Pre-compute nLogN lookup
 	nLogN := make([]float64, k+1)
 	for i := 1; i <= k; i++ {
 		nLogN[i] = float64(i) * math.Log(float64(i))
@@ -51,6 +50,23 @@ func KmerEntropy(kmer uint64, k int, levelMax int) float64 {
 		}
 	}
 
+	// Build ln(class_size) tables: for each canonical form, how many raw
+	// words map to it under circular normalization.
+	classLogSizeTables := make([][]float64, levelMax+1)
+	for ws := 1; ws <= levelMax; ws++ {
+		tableSize := 1 << (ws * 2)
+		classSize := make([]int, tableSize)
+		for code := 0; code < tableSize; code++ {
+			classSize[normTables[ws][code]]++
+		}
+		classLogSizeTables[ws] = make([]float64, tableSize)
+		for j := 0; j < tableSize; j++ {
+			if classSize[j] > 0 {
+				classLogSizeTables[ws][j] = math.Log(float64(classSize[j]))
+			}
+		}
+	}
+
 	minEntropy := math.MaxFloat64
 
 	for ws := 1; ws <= levelMax; ws++ {
@@ -75,23 +91,13 @@ func KmerEntropy(kmer uint64, k int, levelMax int) float64 {
 			table[normWord]++
 		}
 
-		// Compute Shannon entropy
+		// Compute emax over 4^ws raw bins (uncollapsed distribution).
 		floatNwords := float64(nwords)
 		logNwords := math.Log(floatNwords)
-
+		na := tableSize // 4^ws
-		var sumNLogN float64
-		for j := 0; j < tableSize; j++ {
-			n := table[j]
-			if n > 0 {
-				sumNLogN += nLogN[n]
-			}
-		}
-
-		// Compute emax (maximum possible entropy for this word size)
-		na := CanonicalCircularKmerCount(ws)
 		var emax float64
 		if nwords < na {
-			emax = math.Log(float64(nwords))
+			emax = logNwords
 		} else {
 			cov := nwords / na
 			remains := nwords - (na * cov)
@@ -105,7 +111,19 @@ func KmerEntropy(kmer uint64, k int, levelMax int) float64 {
 			continue
 		}
 
-		entropy := (logNwords - sumNLogN/floatNwords) / emax
+		// Accumulate Σ f·log(f) and Σ f·log(class_size) over canonical forms.
+		classLogSize := classLogSizeTables[ws]
+		var sumNLogN, sumClassLogN float64
+		for j := 0; j < tableSize; j++ {
+			n := table[j]
+			if n > 0 {
+				sumNLogN += nLogN[n]
+				sumClassLogN += float64(n) * classLogSize[j]
+			}
+		}
+
+		// Corrected entropy: H_raw ≈ log(N) + (Σf·log(s) - Σf·log(f)) / N
+		entropy := (logNwords + sumClassLogN/floatNwords - sumNLogN/floatNwords) / emax
 		if entropy < 0 {
 			entropy = 0
 		}
@@ -134,6 +152,7 @@ type KmerEntropyFilter struct {
 	threshold           float64
 	nLogN               []float64
 	normTables          [][]int
+	classLogSizeTables  [][]float64
 	emaxValues          []float64
 	logNwords           []float64
 	// Pre-allocated frequency tables reused across Entropy() calls.
@@ -142,11 +161,6 @@ type KmerEntropyFilter struct {
 }
 
 // NewKmerEntropyFilter creates an entropy filter with pre-computed tables.
-//
-// Parameters:
-//   - k: the k-mer size
-//   - levelMax: maximum sub-word size for entropy (typically 6)
-//   - threshold: entropy threshold (k-mers with entropy <= threshold are rejected)
 func NewKmerEntropyFilter(k, levelMax int, threshold float64) *KmerEntropyFilter {
 	if levelMax >= k {
 		levelMax = k - 1
@@ -169,20 +183,38 @@ func NewKmerEntropyFilter(k, levelMax int, threshold float64) *KmerEntropyFilter
 		}
 	}
 
+	// ln(class_size) for each canonical form under circular normalization.
+	classLogSizeTables := make([][]float64, levelMax+1)
+	for ws := 1; ws <= levelMax; ws++ {
+		tableSize := 1 << (ws * 2)
+		classSize := make([]int, tableSize)
+		for code := 0; code < tableSize; code++ {
+			classSize[normTables[ws][code]]++
+		}
+		classLogSizeTables[ws] = make([]float64, tableSize)
+		for j := 0; j < tableSize; j++ {
+			if classSize[j] > 0 {
+				classLogSizeTables[ws][j] = math.Log(float64(classSize[j]))
+			}
+		}
+	}
+
+	// Pre-compute emax and logNwords per word size.
+	// emax uses 4^ws raw bins to match the corrected entropy.
 	emaxValues := make([]float64, levelMax+1)
 	logNwords := make([]float64, levelMax+1)
 	for ws := 1; ws <= levelMax; ws++ {
 		nw := k - ws + 1
-		na := CanonicalCircularKmerCount(ws)
+		na := 1 << (ws * 2) // 4^ws raw bins
+		floatNw := float64(nw)
+		logNwords[ws] = math.Log(floatNw)
 		if nw < na {
-			logNwords[ws] = math.Log(float64(nw))
+			emaxValues[ws] = logNwords[ws]
-			emaxValues[ws] = math.Log(float64(nw))
 		} else {
 			cov := nw / na
 			remains := nw - (na * cov)
-			f1 := float64(cov) / float64(nw)
+			f1 := float64(cov) / floatNw
-			f2 := float64(cov+1) / float64(nw)
+			f2 := float64(cov+1) / floatNw
-			logNwords[ws] = math.Log(float64(nw))
 			emaxValues[ws] = -(float64(na-remains)*f1*math.Log(f1) +
 				float64(remains)*f2*math.Log(f2))
 		}
@@ -200,6 +232,7 @@ func NewKmerEntropyFilter(k, levelMax int, threshold float64) *KmerEntropyFilter
 		threshold:          threshold,
 		nLogN:              nLogN,
 		normTables:         normTables,
+		classLogSizeTables: classLogSizeTables,
 		emaxValues:         emaxValues,
 		logNwords:          logNwords,
 		freqTables:         freqTables,
@@ -236,7 +269,7 @@ func (ef *KmerEntropyFilter) Entropy(kmer uint64) float64 {
 		// Count circular-canonical sub-word frequencies
 		tableSize := 1 << (ws * 2)
 		table := ef.freqTables[ws]
-		clear(table) // reset to zero
+		clear(table)
 		mask := (1 << (ws * 2)) - 1
 		normTable := ef.normTables[ws]
 
@@ -251,19 +284,21 @@ func (ef *KmerEntropyFilter) Entropy(kmer uint64) float64 {
 			table[normWord]++
 		}
 
-		// Compute Shannon entropy
 		floatNwords := float64(nwords)
 		logNwords := ef.logNwords[ws]
+		classLogSize := ef.classLogSizeTables[ws]
 
-		var sumNLogN float64
+		var sumNLogN, sumClassLogN float64
 		for j := 0; j < tableSize; j++ {
 			n := table[j]
 			if n > 0 {
 				sumNLogN += ef.nLogN[n]
+				sumClassLogN += float64(n) * classLogSize[j]
 			}
 		}
 
-		entropy := (logNwords - sumNLogN/floatNwords) / emax
+		// Corrected entropy: H_raw ≈ log(N) + (Σf·log(s) - Σf·log(f)) / N
+		entropy := (logNwords + sumClassLogN/floatNwords - sumNLogN/floatNwords) / emax
 		if entropy < 0 {
 			entropy = 0
 		}

pkg/obioptions/version.go

@@ -3,7 +3,7 @@ package obioptions
 // Version is automatically updated by the Makefile from version.txt
 // The patch number (third digit) is incremented on each push to the repository
 
-var _Version = "Release 4.4.42"
+var _Version = "Release 4.4.46"
 
 // Version returns the version of the obitools package.
 //

pkg/obiseq/attributes.go

@@ -364,6 +364,24 @@ func (s *BioSequence) GetIntSlice(key string) ([]int, bool) {
 	return val, ok
 }
 
+func (s *BioSequence) GetMapOfIntSlice(key string) (map[string][]int, bool) {
+	v, ok := s.GetAttribute(key)
+	if !ok {
+		return nil, false
+	}
+	val, err := obiutils.InterfaceToMapOfIntSlice(v)
+	return val, err == nil
+}
+
+func (s *BioSequence) GetMapOfStringSlice(key string) (map[string][]string, bool) {
+	v, ok := s.GetAttribute(key)
+	if !ok {
+		return nil, false
+	}
+	val, err := obiutils.InterfaceToMapOfStringSlice(v)
+	return val, err == nil
+}
+
 // Count returns the value of the "count" attribute of the BioSequence.
 //
 // The count of a sequence is the number of times it has been observed in the dataset.

pkg/obiseq/biosequence_test.go

@@ -103,7 +103,7 @@ func TestNewBioSequence(t *testing.T) {
 // Return type: None.
 func TestNewBioSequenceWithQualities(t *testing.T) {
 	id := "123"
-	sequence := []byte("ATGC")
+	sequence := []byte("atgc")
 	definition := "DNA sequence"
 	qualities := []byte("1234")
 

pkg/obiseq/language.go

@@ -141,6 +141,33 @@ var OBILang = gval.NewLanguage(
 	gval.Function("max", func(args ...interface{}) (interface{}, error) {
 		return obiutils.Max(args[0])
 	}),
+	gval.Function("whichmax", func(args ...interface{}) (interface{}, error) {
+		result, err := obiutils.WhichMax(args[0])
+		if idx, ok := result.(int); ok {
+			return float64(idx), nil
+		}
+		return result, err
+	}),
+	gval.Function("whichmin", func(args ...interface{}) (interface{}, error) {
+		result, err := obiutils.WhichMin(args[0])
+		if idx, ok := result.(int); ok {
+			return float64(idx), nil
+		}
+		return result, err
+	}),
+
+	gval.Function("filtermin", func(args ...interface{}) (interface{}, error) {
+		return obiutils.FilterMin(args[0], args[1])
+	}),
+
+	gval.Function("filtermax", func(args ...interface{}) (interface{}, error) {
+		return obiutils.FilterMax(args[0], args[1])
+	}),
+
+	gval.Function("saturatingsub", func(args ...interface{}) (interface{}, error) {
+		return obiutils.SaturatingSub(args[0], args[1])
+	}),
+
 	gval.Function("contains", func(args ...interface{}) (interface{}, error) {
 		if obiutils.IsAMap(args[0]) {
 			val := reflect.ValueOf(args[0]).MapIndex(reflect.ValueOf(args[1]))

pkg/obiutils/cast_interface.go

@@ -276,6 +276,44 @@ func InterfaceToStringMap(i interface{}) (val map[string]string, err error) {
 	return
 }
 
+func InterfaceToMapOfIntSlice(i interface{}) (val map[string][]int, err error) {
+	err = nil
+	switch m := i.(type) {
+	case map[string][]int:
+		val = m
+	case map[string]interface{}:
+		val = make(map[string][]int, len(m))
+		for k, v := range m {
+			val[k], err = InterfaceToIntSlice(v)
+			if err != nil {
+				return
+			}
+		}
+	default:
+		err = &NotAMapInt{"value attribute cannot be casted to a map[string][]int"}
+	}
+	return
+}
+
+func InterfaceToMapOfStringSlice(i interface{}) (val map[string][]string, err error) {
+	err = nil
+	switch m := i.(type) {
+	case map[string][]string:
+		val = m
+	case map[string]interface{}:
+		val = make(map[string][]string, len(m))
+		for k, v := range m {
+			val[k], err = InterfaceToStringSlice(v)
+			if err != nil {
+				return
+			}
+		}
+	default:
+		err = &NotAMapInt{"value attribute cannot be casted to a map[string][]string"}
+	}
+	return
+}
+
 func InterfaceToStringSlice(i interface{}) (val []string, err error) {
 	err = nil
 

pkg/obiutils/minmax.go

@@ -34,6 +34,26 @@ func MinMaxSlice[T constraints.Ordered](vec []T) (min, max T) {
 	return
 }
 
+func FilterMinSlice[T constraints.Ordered](vec []T, minimum T) []T {
+	result := make([]T, 0, len(vec))
+	for _, v := range vec {
+		if v >= minimum {
+			result = append(result, v)
+		}
+	}
+	return result
+}
+
+func FilterMaxSlice[T constraints.Ordered](vec []T, maximum T) []T {
+	result := make([]T, 0, len(vec))
+	for _, v := range vec {
+		if v <= maximum {
+			result = append(result, v)
+		}
+	}
+	return result
+}
+
 func MaxMap[K comparable, T constraints.Ordered](values map[K]T) (K, T, error) {
 	var maxKey K
 	var maxValue T
@@ -73,6 +93,46 @@ func MinMap[K comparable, T constraints.Ordered](values map[K]T) (K, T, error) {
 	return minKey, minValue, nil
 }
 
+func FilterMinMap[K comparable, T constraints.Ordered](values map[K]T, minimum T) map[K]T {
+	result := make(map[K]T)
+	for k, v := range values {
+		if v >= minimum {
+			result[k] = v
+		}
+	}
+	return result
+}
+
+func FilterMaxMap[K comparable, T constraints.Ordered](values map[K]T, maximum T) map[K]T {
+	result := make(map[K]T)
+	for k, v := range values {
+		if v <= maximum {
+			result[k] = v
+		}
+	}
+	return result
+}
+
+func SaturatingSubSlice[T Numeric](vec []T, sub T) []T {
+	result := make([]T, len(vec))
+	for i, v := range vec {
+		if v > sub {
+			result[i] = v - sub
+		}
+	}
+	return result
+}
+
+func SaturatingSubMap[K comparable, T Numeric](values map[K]T, sub T) map[K]T {
+	result := make(map[K]T)
+	for k, v := range values {
+		if v > sub {
+			result[k] = v - sub
+		}
+	}
+	return result
+}
+
 // Min returns the smallest element in a slice/array or map,
 // or the value itself if data is a single comparable value.
 // Returns an error if the container is empty or the type is unsupported.
@@ -135,11 +195,121 @@ func Max(data interface{}) (interface{}, error) {
 	}
 }
 
+func FilterMin(data interface{}, minimum interface{}) (interface{}, error) {
+	v := reflect.ValueOf(data)
+	switch v.Kind() {
+	case reflect.Slice, reflect.Array:
+		if v.Len() == 0 {
+			return nil, errors.New("empty slice or array")
+		}
+		return filterMinFromIterable(v, minimum)
+	case reflect.Map:
+		if v.Len() == 0 {
+			return nil, errors.New("empty map")
+		}
+		return filterMinFromMap(v, minimum)
+	default:
+		if !isOrderedKind(v.Kind()) {
+			return nil, fmt.Errorf("unsupported type: %s", v.Kind())
+		}
+		return data, nil
+	}
+}
+
+func FilterMax(data interface{}, maximum interface{}) (interface{}, error) {
+	v := reflect.ValueOf(data)
+	switch v.Kind() {
+	case reflect.Slice, reflect.Array:
+		if v.Len() == 0 {
+			return nil, errors.New("empty slice or array")
+		}
+		return filterMaxFromIterable(v, maximum)
+	case reflect.Map:
+		if v.Len() == 0 {
+			return nil, errors.New("empty map")
+		}
+		return filterMaxFromMap(v, maximum)
+	default:
+		if !isOrderedKind(v.Kind()) {
+			return nil, fmt.Errorf("unsupported type: %s", v.Kind())
+		}
+		return data, nil
+	}
+}
+
+func SaturatingSub(data interface{}, sub interface{}) (interface{}, error) {
+	v := reflect.ValueOf(data)
+	switch v.Kind() {
+	case reflect.Slice, reflect.Array:
+		return saturatingSubFromIterable(v, sub)
+	case reflect.Map:
+		return saturatingSubFromMap(v, sub)
+	default:
+		if !isNumericKind(v.Kind()) {
+			return nil, fmt.Errorf("unsupported type: %s", v.Kind())
+		}
+		r, err := saturatingSubValues(v, reflect.ValueOf(sub))
+		if err != nil {
+			return nil, err
+		}
+		return r.Interface(), nil
+	}
+}
+
+func saturatingSubFromIterable(v reflect.Value, sub interface{}) (interface{}, error) {
+	subVal := reflect.ValueOf(sub)
+	result := reflect.MakeSlice(v.Type(), v.Len(), v.Len())
+	for i := 0; i < v.Len(); i++ {
+		r, err := saturatingSubValues(v.Index(i), subVal)
+		if err != nil {
+			return nil, err
+		}
+		result.Index(i).Set(r)
+	}
+	return result.Interface(), nil
+}
+
+func saturatingSubFromMap(v reflect.Value, sub interface{}) (interface{}, error) {
+	subVal := reflect.ValueOf(sub)
+	result := reflect.MakeMap(v.Type())
+	for _, key := range v.MapKeys() {
+		r, err := saturatingSubValues(v.MapIndex(key), subVal)
+		if err != nil {
+			return nil, err
+		}
+		if !r.IsZero() {
+			result.SetMapIndex(key, r)
+		}
+	}
+	return result.Interface(), nil
+}
+
+func saturatingSubValues(a, b reflect.Value) (reflect.Value, error) {
+	result := reflect.New(a.Type()).Elem()
+	switch a.Kind() {
+	case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
+		if av, bv := a.Int(), b.Int(); av > bv {
+			result.SetInt(av - bv)
+		}
+	case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64:
+		if av, bv := a.Uint(), b.Uint(); av > bv {
+			result.SetUint(av - bv)
+		}
+	case reflect.Float32, reflect.Float64:
+		if av, bv := a.Float(), b.Float(); av > bv {
+			result.SetFloat(av - bv)
+		}
+	default:
+		return reflect.Value{}, fmt.Errorf("unsupported type for saturating subtraction: %s", a.Kind())
+	}
+	return result, nil
+}
+
 // maxFromIterable scans a slice/array to find the maximum.
 func maxFromIterable(v reflect.Value) (interface{}, error) {
 	var best reflect.Value
 	for i := 0; i < v.Len(); i++ {
-		elem := v.Index(i)
+		elem := unwrapInterface(v.Index(i))
 		if !isOrderedKind(elem.Kind()) {
 			return nil, fmt.Errorf("unsupported element type: %s", elem.Kind())
 		}
@@ -154,7 +324,7 @@ func maxFromIterable(v reflect.Value) (interface{}, error) {
 func minFromIterable(v reflect.Value) (interface{}, error) {
 	var minVal reflect.Value
 	for i := 0; i < v.Len(); i++ {
-		elem := v.Index(i)
+		elem := unwrapInterface(v.Index(i))
 		if !isOrderedKind(elem.Kind()) {
 			return nil, fmt.Errorf("unsupported element type: %s", elem.Kind())
 		}
@@ -165,12 +335,182 @@ func minFromIterable(v reflect.Value) (interface{}, error) {
 	return minVal.Interface(), nil
 }
 
+func filterMinFromIterable(v reflect.Value, minimum interface{}) (interface{}, error) {
+	minVal := reflect.ValueOf(minimum)
+	result := reflect.MakeSlice(v.Type(), 0, v.Len())
+	for i := 0; i < v.Len(); i++ {
+		elem := unwrapInterface(v.Index(i))
+		if !isOrderedKind(elem.Kind()) {
+			return nil, fmt.Errorf("unsupported element type: %s", elem.Kind())
+		}
+		if !less(elem, minVal) { // elem >= minimum
+			result = reflect.Append(result, elem)
+		}
+	}
+	return result.Interface(), nil
+}
+
+func filterMaxFromIterable(v reflect.Value, maximum interface{}) (interface{}, error) {
+	maxVal := reflect.ValueOf(maximum)
+	result := reflect.MakeSlice(v.Type(), 0, v.Len())
+	for i := 0; i < v.Len(); i++ {
+		elem := unwrapInterface(v.Index(i))
+		if !isOrderedKind(elem.Kind()) {
+			return nil, fmt.Errorf("unsupported element type: %s", elem.Kind())
+		}
+		if !greater(elem, maxVal) { // elem <= maximum
+			result = reflect.Append(result, elem)
+		}
+	}
+	return result.Interface(), nil
+}
+
+// whichMaxFromIterable returns the index of the maximum element in a slice/array.
+func whichMaxFromIterable(v reflect.Value) (int, error) {
+	var best reflect.Value
+	bestIdx := 0
+	for i := 0; i < v.Len(); i++ {
+		elem := unwrapInterface(v.Index(i))
+		if !isOrderedKind(elem.Kind()) {
+			return 0, fmt.Errorf("unsupported element type: %s", elem.Kind())
+		}
+		if i == 0 || greater(elem, best) {
+			best = elem
+			bestIdx = i
+		}
+	}
+	return bestIdx, nil
+}
+
+// whichMinFromIterable returns the index of the minimum element in a slice/array.
+func whichMinFromIterable(v reflect.Value) (int, error) {
+	var minVal reflect.Value
+	minIdx := 0
+	for i := 0; i < v.Len(); i++ {
+		elem := unwrapInterface(v.Index(i))
+		if !isOrderedKind(elem.Kind()) {
+			return 0, fmt.Errorf("unsupported element type: %s", elem.Kind())
+		}
+		if i == 0 || less(elem, minVal) {
+			minVal = elem
+			minIdx = i
+		}
+	}
+	return minIdx, nil
+}
+
+// whichMaxFromMap returns the key associated with the maximum value in a map.
+func whichMaxFromMap(v reflect.Value) (interface{}, error) {
+	var best reflect.Value
+	var bestKey reflect.Value
+	first := true
+	for _, key := range v.MapKeys() {
+		elem := unwrapInterface(v.MapIndex(key))
+		if !isOrderedKind(elem.Kind()) {
+			return nil, fmt.Errorf("unsupported element type: %s", elem.Kind())
+		}
+		if first || greater(elem, best) {
+			best = elem
+			bestKey = key
+			first = false
+		}
+	}
+	return bestKey.Interface(), nil
+}
+
+// whichMinFromMap returns the key associated with the minimum value in a map.
+func whichMinFromMap(v reflect.Value) (interface{}, error) {
+	var minVal reflect.Value
+	var minKey reflect.Value
+	first := true
+	for _, key := range v.MapKeys() {
+		elem := unwrapInterface(v.MapIndex(key))
+		if !isOrderedKind(elem.Kind()) {
+			return nil, fmt.Errorf("unsupported element type: %s", elem.Kind())
+		}
+		if first || less(elem, minVal) {
+			minVal = elem
+			minKey = key
+			first = false
+		}
+	}
+	return minKey.Interface(), nil
+}
+
+// WhichMax returns the key (for a map) or index (for a slice/array) of the maximum value.
+func WhichMax(data interface{}) (interface{}, error) {
+	v := reflect.ValueOf(data)
+	switch v.Kind() {
+	case reflect.Slice, reflect.Array:
+		if v.Len() == 0 {
+			return nil, errors.New("empty slice or array")
+		}
+		return whichMaxFromIterable(v)
+	case reflect.Map:
+		if v.Len() == 0 {
+			return nil, errors.New("empty map")
+		}
+		return whichMaxFromMap(v)
+	default:
+		return nil, fmt.Errorf("unsupported type: %s", v.Kind())
+	}
+}
+
+// WhichMin returns the key (for a map) or index (for a slice/array) of the minimum value.
+func WhichMin(data interface{}) (interface{}, error) {
+	v := reflect.ValueOf(data)
+	switch v.Kind() {
+	case reflect.Slice, reflect.Array:
+		if v.Len() == 0 {
+			return nil, errors.New("empty slice or array")
+		}
+		return whichMinFromIterable(v)
+	case reflect.Map:
+		if v.Len() == 0 {
+			return nil, errors.New("empty map")
+		}
+		return whichMinFromMap(v)
+	default:
+		return nil, fmt.Errorf("unsupported type: %s", v.Kind())
+	}
+}
+
+func filterMinFromMap(v reflect.Value, minimum interface{}) (interface{}, error) {
+	minVal := reflect.ValueOf(minimum)
+	result := reflect.MakeMap(v.Type())
+	for _, key := range v.MapKeys() {
+		elem := unwrapInterface(v.MapIndex(key))
+		if !isOrderedKind(elem.Kind()) {
+			return nil, fmt.Errorf("unsupported element type: %s", elem.Kind())
+		}
+		if !less(elem, minVal) { // elem >= minimum
+			result.SetMapIndex(key, elem)
+		}
+	}
+	return result.Interface(), nil
+}
+
+func filterMaxFromMap(v reflect.Value, maximum interface{}) (interface{}, error) {
+	maxVal := reflect.ValueOf(maximum)
+	result := reflect.MakeMap(v.Type())
+	for _, key := range v.MapKeys() {
+		elem := unwrapInterface(v.MapIndex(key))
+		if !isOrderedKind(elem.Kind()) {
+			return nil, fmt.Errorf("unsupported element type: %s", elem.Kind())
+		}
+		if !greater(elem, maxVal) { // elem <= maximum
+			result.SetMapIndex(key, elem)
+		}
+	}
+	return result.Interface(), nil
+}
+
 // maxFromMap scans map values to find the maximum.
 func maxFromMap(v reflect.Value) (interface{}, error) {
 	var best reflect.Value
 	first := true
 	for _, key := range v.MapKeys() {
-		elem := v.MapIndex(key)
+		elem := unwrapInterface(v.MapIndex(key))
 		if !isOrderedKind(elem.Kind()) {
 			return nil, fmt.Errorf("unsupported element type: %s", elem.Kind())
 		}
@@ -187,7 +527,7 @@ func minFromMap(v reflect.Value) (interface{}, error) {
 	var minVal reflect.Value
 	first := true
 	for _, key := range v.MapKeys() {
-		elem := v.MapIndex(key)
+		elem := unwrapInterface(v.MapIndex(key))
 		if !isOrderedKind(elem.Kind()) {
 			return nil, fmt.Errorf("unsupported element type: %s", elem.Kind())
 		}
@@ -199,6 +539,27 @@ func minFromMap(v reflect.Value) (interface{}, error) {
 	return minVal.Interface(), nil
 }
 
+func isNumericKind(k reflect.Kind) bool {
+	switch k {
+	case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64,
+		reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64,
+		reflect.Float32, reflect.Float64:
+		return true
+	default:
+		return false
+	}
+}
+
+// unwrapInterface returns v.Elem() when v holds an interface value, otherwise v unchanged.
+// This is necessary when iterating map[string]interface{} or []interface{} via reflection:
+// the element Kind is reflect.Interface, not the underlying concrete type.
+func unwrapInterface(v reflect.Value) reflect.Value {
+	if v.Kind() == reflect.Interface {
+		return v.Elem()
+	}
+	return v
+}
+
 // isOrderedKind reports whether k supports comparison ordering.
 func isOrderedKind(k reflect.Kind) bool {
 	switch k {

version.txt

@@ -1 +1 @@
-4.4.42
+4.4.46