4.4.25: Static Linux Builds, Memory-Aware Batching, and Build Stability

### Static Linux Builds
- Added `CGO_CFLAGS` environment variable to the release workflow for consistent static linking on Linux, ensuring portable, self-contained executables.
- Updated `go.work.sum` with the new dependency `golang.org/x/net v0.38.0`.
- Removed obsolete logs archive file from the distribution.

### Memory-Aware Batching (Introduced in 4.4.23, now stable)
- Users can now control batching behavior using `--batch-mem` to specify memory limits (e.g., `128K`, `64M`, `1G`), enabling adaptive batching based on sequence data size.
- Batching now respects both byte and record count limits, flushing batches when either threshold is exceeded.
- Conservative memory estimation for sequences (`BioSequence.MemorySize()`) and explicit garbage collection after large batch discards improve resource predictability.
- Default constraints remain non-breaking: minimum 1, maximum 2000 records per batch with a default memory limit of 128 MB.

### Build System Improvements
- Updated Go toolchain to 1.26.1 and bumped key dependencies for security and performance.
- Fixed Makefile quoting for `LDFLAGS` to safely handle paths containing spaces.
- Enhanced build failure handling: error logs are now displayed before cleanup, aiding diagnostics.
- The install script now correctly configures `GOROOT`, `GOPATH`, and `GOTOOLCHAIN`, creates the GOPATH directory, and shows a progress bar during downloads.

.github/workflows/obitest.yml

@@ -10,10 +10,10 @@ jobs:
     runs-on: ubuntu-latest
     steps:
     - name: Setup Go
-      uses: actions/setup-go@v5
+      uses: actions/setup-go@v2
       with:
         go-version: '1.23'
     - name: Checkout obitools4 project
-      uses: actions/checkout@v5
+      uses: actions/checkout@v4
     - name: Run tests
       run: make githubtests

.github/workflows/release.yml

@@ -16,9 +16,9 @@ jobs:
       - name: Setup Go
         uses: actions/setup-go@v5
         with:
-          go-version: "1.26"
+          go-version: "1.23"
       - name: Checkout obitools4 project
-        uses: actions/checkout@v5
+        uses: actions/checkout@v4
       - name: Run tests
         run: make githubtests
 
@@ -49,12 +49,12 @@ jobs:
 
     steps:
       - name: Checkout code
-        uses: actions/checkout@v5
+        uses: actions/checkout@v4
 
       - name: Setup Go
         uses: actions/setup-go@v5
         with:
-          go-version: "1.26"
+          go-version: "1.23"
 
       - name: Extract version from tag
         id: get_version
@@ -62,6 +62,12 @@ jobs:
           TAG=${GITHUB_REF#refs/tags/Release_}
           echo "version=$TAG" >> $GITHUB_OUTPUT
 
+      - name: Install build tools (Linux)
+        if: runner.os == 'Linux'
+        run: |
+          sudo apt-get update -q
+          sudo apt-get install -y musl-tools zlib1g-dev
+
       - name: Install build tools (macOS)
         if: runner.os == 'macOS'
         run: |
@@ -69,30 +75,21 @@ jobs:
           xcode-select --install 2>/dev/null || true
           xcode-select -p
 
-      - name: Build binaries (Linux)
-        if: runner.os == 'Linux'
-        env:
-          VERSION: ${{ steps.get_version.outputs.version }}
-        run: |
-          docker run --rm \
-            -v "$(pwd):/src" \
-            -w /src \
-            -e VERSION="${VERSION}" \
-            golang:1.26-alpine \
-            sh -c "apk add --no-cache gcc musl-dev zlib-dev zlib-static make && \
-                   make LDFLAGS='-linkmode=external -extldflags=-static' obitools"
-          mkdir -p artifacts
-          tar -czf artifacts/obitools4_${VERSION}_${{ matrix.output_name }}.tar.gz -C build .
-
-      - name: Build binaries (macOS)
-        if: runner.os == 'macOS'
+      - name: Build binaries
         env:
           GOOS: ${{ matrix.goos }}
           GOARCH: ${{ matrix.goarch }}
           VERSION: ${{ steps.get_version.outputs.version }}
+          CC: ${{ matrix.goos == 'linux' && 'musl-gcc' || '' }}
+          CGO_CFLAGS: ${{ matrix.goos == 'linux' && '-I/usr/include' || '' }}
         run: |
+          if [ "$GOOS" = "linux" ]; then
+            make LDFLAGS='-linkmode=external -extldflags=-static' obitools
+          else
             make obitools
+          fi
           mkdir -p artifacts
+          # Create a single tar.gz with all binaries for this platform
           tar -czf artifacts/obitools4_${VERSION}_${{ matrix.output_name }}.tar.gz -C build .
 
       - name: Upload artifacts
@@ -107,7 +104,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout code
-        uses: actions/checkout@v5
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 

.gitignore

@@ -23,7 +23,7 @@ xx
 /.vscode
 /build
 /bugs
-autodoc
+
 /ncbitaxo
 
 !/obitests/**

entities.json

@@ -1,10 +0,0 @@
-{
-  "people": [
-    "Software",
-    "Agreement",
-    "Module"
-  ],
-  "projects": [
-    "Code"
-  ]
-}

go.mod

@@ -6,7 +6,7 @@ require (
 	github.com/DavidGamba/go-getoptions v0.33.0
 	github.com/PaesslerAG/gval v1.2.4
 	github.com/barkimedes/go-deepcopy v0.0.0-20220514131651-17c30cfc62df
-	github.com/buger/jsonparser v1.1.2
+	github.com/buger/jsonparser v1.1.1
 	github.com/chen3feng/stl4go v0.1.1
 	github.com/dlclark/regexp2 v1.11.5
 	github.com/goccy/go-json v0.10.6

go.sum

@@ -6,8 +6,8 @@ github.com/PaesslerAG/jsonpath v0.1.0 h1:gADYeifvlqK3R3i2cR5B4DGgxLXIPb3TRTH1mGi
 github.com/PaesslerAG/jsonpath v0.1.0/go.mod h1:4BzmtoM/PI8fPO4aQGIusjGxGir2BzcV0grWtFzq1Y8=
 github.com/barkimedes/go-deepcopy v0.0.0-20220514131651-17c30cfc62df h1:GSoSVRLoBaFpOOds6QyY1L8AX7uoY+Ln3BHc22W40X0=
 github.com/barkimedes/go-deepcopy v0.0.0-20220514131651-17c30cfc62df/go.mod h1:hiVxq5OP2bUGBRNS3Z/bt/reCLFNbdcST6gISi1fiOM=
-github.com/buger/jsonparser v1.1.2 h1:frqHqw7otoVbk5M8LlE/L7HTnIq2v9RX6EJ48i9AxJk=
-github.com/buger/jsonparser v1.1.2/go.mod h1:6RYKKt7H4d4+iWqouImQ9R2FZql3VbhNgx27UK13J/0=
+github.com/buger/jsonparser v1.1.1 h1:2PnMjfWD7wBILjqQbt530v576A/cAbQvEW9gGIpYMUs=
+github.com/buger/jsonparser v1.1.1/go.mod h1:6RYKKt7H4d4+iWqouImQ9R2FZql3VbhNgx27UK13J/0=
 github.com/chen3feng/stl4go v0.1.1 h1:0L1+mDw7pomftKDruM23f1mA7miavOj6C6MZeadzN2Q=
 github.com/chen3feng/stl4go v0.1.1/go.mod h1:5ml3psLgETJjRJnMbPE+JiHLrCpt+Ajc2weeTECXzWU=
 github.com/chengxilo/virtualterm v1.0.4 h1:Z6IpERbRVlfB8WkOmtbHiDbBANU7cimRIof7mk9/PwM=

pkg/obiiter/distribute.go

@@ -57,21 +57,34 @@ func (dist *IDistribute) Classifier() *obiseq.BioSequenceClassifier {
 }
 
 // Distribute organizes the biosequences from the iterator into batches
-// based on the provided classifier. It returns an IDistribute instance
-// that manages the distribution of the sequences.
+// based on the provided classifier and batch sizes. It returns an
+// IDistribute instance that manages the distribution of the sequences.
 //
-// Batches are flushed when either BatchSizeMax() sequences or BatchMem()
-// bytes are accumulated per key, mirroring the RebatchBySize strategy.
-func (iterator IBioSequence) Distribute(class *obiseq.BioSequenceClassifier) IDistribute {
-	maxCount := obidefault.BatchSizeMax()
-	maxBytes := obidefault.BatchMem()
+// Parameters:
+//   - class: A pointer to a BioSequenceClassifier used to classify
+//     the biosequences during distribution.
+//   - sizes: Optional integer values specifying the batch size. If
+//     no sizes are provided, a default batch size of 5000 is used.
+//
+// Returns:
+// An IDistribute instance that contains the outputs of the
+// classified biosequences, a channel for new data notifications,
+// and the classifier used for distribution. The method operates
+// asynchronously, processing the sequences in separate goroutines.
+// It ensures that the outputs are closed and cleaned up once
+// processing is complete.
+func (iterator IBioSequence) Distribute(class *obiseq.BioSequenceClassifier, sizes ...int) IDistribute {
+	batchsize := obidefault.BatchSize()
 
 	outputs := make(map[int]IBioSequence, 100)
 	slices := make(map[int]*obiseq.BioSequenceSlice, 100)
-	bufBytes := make(map[int]int, 100)
 	orders := make(map[int]int, 100)
 	news := make(chan int)
 
+	if len(sizes) > 0 {
+		batchsize = sizes[0]
+	}
+
 	jobDone := sync.WaitGroup{}
 	lock := sync.Mutex{}
 
@@ -102,7 +115,6 @@ func (iterator IBioSequence) Distribute(class *obiseq.BioSequenceClassifier) IDi
 					slice = &s
 					slices[key] = slice
 					orders[key] = 0
-					bufBytes[key] = 0
 
 					lock.Lock()
 					outputs[key] = MakeIBioSequence()
@@ -111,20 +123,14 @@ func (iterator IBioSequence) Distribute(class *obiseq.BioSequenceClassifier) IDi
 					news <- key
 				}
 
-				sz := s.MemorySize()
-				countFull := maxCount > 0 && len(*slice) >= maxCount
-				memFull := maxBytes > 0 && bufBytes[key]+sz > maxBytes && len(*slice) > 0
-				if countFull || memFull {
+				*slice = append(*slice, s)
+
+				if len(*slice) == batchsize {
 					outputs[key].Push(MakeBioSequenceBatch(source, orders[key], *slice))
 					orders[key]++
 					s := obiseq.MakeBioSequenceSlice()
 					slices[key] = &s
-					slice = &s
-					bufBytes[key] = 0
 				}
-
-				*slice = append(*slice, s)
-				bufBytes[key] += sz
 			}
 		}
 

pkg/obikmer/encodefourmer.go

@@ -47,7 +47,7 @@ func Encode4mer(seq *obiseq.BioSequence, buffer *[]byte) []byte {
 	length := slength - 3
 	rawseq := seq.Sequence()
 
-	if length <= 0 {
+	if length < 0 {
 		return nil
 	}
 

pkg/obilua/lua.go

@@ -91,7 +91,7 @@ func LuaWorker(proto *lua.FunctionProto) obiseq.SeqWorker {
 	err := interpreter.PCall(0, lua.MultRet, nil)
 
 	if err != nil {
-		log.Fatalf("Error in executing the lua script: %v", err)
+		log.Fatalf("Error in executing the lua script")
 	}
 
 	result := interpreter.GetGlobal("worker")
@@ -141,69 +141,6 @@ func LuaWorker(proto *lua.FunctionProto) obiseq.SeqWorker {
 	return nil
 }
 
-// LuaSliceWorker creates a SeqSliceWorker that calls the Lua function
-// named "slice_worker". Unlike LuaWorker, the entire batch (BioSequenceSlice)
-// is passed to the Lua function at once, enabling batch-level processing
-// (e.g. a single HTTP request per batch instead of one per sequence).
-//
-// The Lua function signature:
-//
-//	function slice_worker(slice)   -- receives a BioSequenceSlice
-//	    -- process the batch
-//	    return slice               -- returns a BioSequenceSlice (or nil)
-//	end
-func LuaSliceWorker(proto *lua.FunctionProto) obiseq.SeqSliceWorker {
-	interpreter := NewInterpreter()
-	lfunc := interpreter.NewFunctionFromProto(proto)
-	interpreter.Push(lfunc)
-	err := interpreter.PCall(0, lua.MultRet, nil)
-
-	if err != nil {
-		log.Fatalf("Error in executing the lua script: %v", err)
-	}
-
-	result := interpreter.GetGlobal("slice_worker")
-
-	if lua_worker, ok := result.(*lua.LFunction); ok {
-		f := func(slice obiseq.BioSequenceSlice) (obiseq.BioSequenceSlice, error) {
-			if err := interpreter.CallByParam(lua.P{
-				Fn:      lua_worker,
-				NRet:    1,
-				Protect: true,
-			}, obiseqslice2Lua(interpreter, &slice)); err != nil {
-				log.Fatal(err)
-			}
-
-			lreponse := interpreter.Get(-1)
-			defer interpreter.Pop(1)
-
-			if reponse, ok := lreponse.(*lua.LUserData); ok {
-				s := reponse.Value
-				switch val := s.(type) {
-				case *obiseq.BioSequenceSlice:
-					return *val, nil
-				case *obiseq.BioSequence:
-					return obiseq.BioSequenceSlice{val}, nil
-				default:
-					r := reflect.TypeOf(val)
-					return nil, fmt.Errorf("slice_worker function doesn't return the correct type %s", r)
-				}
-			}
-
-			if _, ok = lreponse.(*lua.LNilType); ok {
-				return nil, nil
-			}
-
-			return nil, fmt.Errorf("slice_worker function doesn't return the correct type %T", lreponse)
-		}
-
-		return f
-	}
-
-	log.Fatalf("The slice_worker object is not a function")
-	return nil
-}
-
 // LuaProcessor processes a Lua script on a sequence iterator and returns a new iterator.
 //
 // Parameters:
@@ -236,7 +173,7 @@ func LuaProcessor(iterator obiiter.IBioSequence, name, program string, breakOnEr
 	err = interpreter.PCall(0, lua.MultRet, nil)
 
 	if err != nil {
-		log.Fatalf("Error in executing the lua script: %v", err)
+		log.Fatalf("Error in executing the lua script")
 	}
 
 	result := interpreter.GetGlobal("begin")
@@ -261,7 +198,7 @@ func LuaProcessor(iterator obiiter.IBioSequence, name, program string, breakOnEr
 		err = interpreter.PCall(0, lua.MultRet, nil)
 
 		if err != nil {
-			log.Fatalf("Error in executing the lua script: %v", err)
+			log.Fatalf("Error in executing the lua script")
 		}
 
 		result := interpreter.GetGlobal("finish")
@@ -279,27 +216,11 @@ func LuaProcessor(iterator obiiter.IBioSequence, name, program string, breakOnEr
 
 	}()
 
-	// Detect whether the script defines slice_worker (batch-level) or worker (per-sequence).
-	hasSliceWorker := func() bool {
-		interpreter := NewInterpreter()
-		lfunc := interpreter.NewFunctionFromProto(proto)
-		interpreter.Push(lfunc)
-		if err := interpreter.PCall(0, lua.MultRet, nil); err != nil {
-			return false
-		}
-		result := interpreter.GetGlobal("slice_worker")
-		interpreter.Close()
-		_, ok := result.(*lua.LFunction)
-		return ok
-	}()
-
 	ff := func(iterator obiiter.IBioSequence) {
-		var sw obiseq.SeqSliceWorker
-		if hasSliceWorker {
-			sw = LuaSliceWorker(proto)
-		} else {
-			sw = obiseq.SeqToSliceWorker(LuaWorker(proto), false)
-		}
+		w := LuaWorker(proto)
+		sw := obiseq.SeqToSliceWorker(w, false)
+
+		// iterator = iterator.SortBatches()
 
 		for iterator.Next() {
 			seqs := iterator.Get()
@@ -314,10 +235,6 @@ func LuaProcessor(iterator obiiter.IBioSequence, name, program string, breakOnEr
 				}
 			}
 
-			if ns == nil {
-				ns = obiseq.BioSequenceSlice{}
-			}
-
 			newIter.Push(obiiter.MakeBioSequenceBatch(seqs.Source(), seqs.Order(), ns))
 		}
 

pkg/obilua/lua_push_interface.go

@@ -17,7 +17,15 @@ import (
 // No return values. This function operates directly on the Lua state stack.
 func pushInterfaceToLua(L *lua.LState, val interface{}) {
 	switch v := val.(type) {
-	// Typed slices and maps from internal OBITools code — not produced by json.Unmarshal
+	case string:
+		L.Push(lua.LString(v))
+	case bool:
+		L.Push(lua.LBool(v))
+	case int:
+		L.Push(lua.LNumber(v))
+	case float64:
+		L.Push(lua.LNumber(v))
+	// Add other cases as needed for different types
 	case map[string]int:
 		pushMapStringIntToLua(L, v)
 	case map[string]string:
@@ -26,6 +34,8 @@ func pushInterfaceToLua(L *lua.LState, val interface{}) {
 		pushMapStringBoolToLua(L, v)
 	case map[string]float64:
 		pushMapStringFloat64ToLua(L, v)
+	case map[string]interface{}:
+		pushMapStringInterfaceToLua(L, v)
 	case []string:
 		pushSliceStringToLua(L, v)
 	case []int:
@@ -36,61 +46,61 @@ func pushInterfaceToLua(L *lua.LState, val interface{}) {
 		pushSliceNumericToLua(L, v)
 	case []bool:
 		pushSliceBoolToLua(L, v)
+	case []interface{}:
+		pushSliceInterfaceToLua(L, v)
+	case nil:
+		L.Push(lua.LNil)
 	case *sync.Mutex:
 		pushMutexToLua(L, v)
 	default:
-		// Handles nil, bool, int, float64, string, map[string]interface{},
-		// []interface{} — all recursively via lvalueFromInterface.
-		L.Push(lvalueFromInterface(L, v))
+		log.Fatalf("Cannot deal with value (%T) : %v", val, val)
 	}
 }
 
 func pushMapStringInterfaceToLua(L *lua.LState, m map[string]interface{}) {
+	// Create a new Lua table
 	luaTable := L.NewTable()
+	// Iterate over the Go map and set the key-value pairs in the Lua table
 	for key, value := range m {
-		L.SetField(luaTable, key, lvalueFromInterface(L, value))
+		switch v := value.(type) {
+		case int:
+			luaTable.RawSetString(key, lua.LNumber(v))
+		case float64:
+			luaTable.RawSetString(key, lua.LNumber(v))
+		case bool:
+			luaTable.RawSetString(key, lua.LBool(v))
+		case string:
+			luaTable.RawSetString(key, lua.LString(v))
+		default:
+			log.Fatalf("Doesn't deal with map containing value %v of type %T", v, v)
 		}
+	}
+
+	// Push the Lua table onto the stack
 	L.Push(luaTable)
 }
 
 func pushSliceInterfaceToLua(L *lua.LState, s []interface{}) {
+	// Create a new Lua table
 	luaTable := L.NewTable()
+	// Iterate over the Go map and set the key-value pairs in the Lua table
 	for _, value := range s {
-		luaTable.Append(lvalueFromInterface(L, value))
-	}
-	L.Push(luaTable)
-}
-
-// lvalueFromInterface converts a Go interface{} value (as produced by json.Unmarshal)
-// to the corresponding lua.LValue, handling nested maps and slices recursively.
-func lvalueFromInterface(L *lua.LState, value interface{}) lua.LValue {
 		switch v := value.(type) {
-	case nil:
-		return lua.LNil
-	case bool:
-		return lua.LBool(v)
 		case int:
-		return lua.LNumber(v)
+			luaTable.Append(lua.LNumber(v))
 		case float64:
-		return lua.LNumber(v)
+			luaTable.Append(lua.LNumber(v))
+		case bool:
+			luaTable.Append(lua.LBool(v))
 		case string:
-		return lua.LString(v)
-	case map[string]interface{}:
-		t := L.NewTable()
-		for key, val := range v {
-			L.SetField(t, key, lvalueFromInterface(L, val))
-		}
-		return t
-	case []interface{}:
-		t := L.NewTable()
-		for _, val := range v {
-			t.Append(lvalueFromInterface(L, val))
-		}
-		return t
+			luaTable.Append(lua.LString(v))
 		default:
-		log.Fatalf("lvalueFromInterface: unsupported type %T: %v", v, v)
-		return lua.LNil
+			log.Fatalf("Doesn't deal with slice containing value %v of type %T", v, v)
 		}
+	}
+
+	// Push the Lua table onto the stack
+	L.Push(luaTable)
 }
 
 // pushMapStringIntToLua creates a new Lua table and iterates over the Go map to set key-value pairs in the Lua table. It then pushes the Lua table onto the stack.

pkg/obilua/lua_table.go

@@ -28,8 +28,6 @@ func Table2Interface(interpreter *lua.LState, table *lua.LTable) interface{} {
 				val[i-1] = float64(v.(lua.LNumber))
 			case lua.LTString:
 				val[i-1] = string(v.(lua.LString))
-			case lua.LTTable:
-				val[i-1] = Table2Interface(interpreter, v.(*lua.LTable))
 			}
 		}
 		return val
@@ -47,8 +45,6 @@ func Table2Interface(interpreter *lua.LState, table *lua.LTable) interface{} {
 					val[string(ks)] = float64(v.(lua.LNumber))
 				case lua.LTString:
 					val[string(ks)] = string(v.(lua.LString))
-				case lua.LTTable:
-					val[string(ks)] = Table2Interface(interpreter, v.(*lua.LTable))
 				}
 			}
 		})

pkg/obilua/luahttp.go

@@ -1,128 +0,0 @@
-package obilua
-
-import (
-	"context"
-	"io"
-	"net/http"
-	"strings"
-	"sync"
-	"time"
-
-	"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obidefault"
-	lua "github.com/yuin/gopher-lua"
-)
-
-const httpClientTimeout = 300 * time.Second
-
-var (
-	_httpClient     *http.Client
-	_httpClientOnce sync.Once
-
-	// _httpSemaphore limits the number of concurrent HTTP requests.
-	// Initialised lazily alongside the client.
-	_httpSemaphore chan struct{}
-)
-
-func getHTTPClient() *http.Client {
-	_httpClientOnce.Do(func() {
-		conns := 2 * obidefault.ParallelWorkers()
-		_httpClient = &http.Client{
-			Transport: &http.Transport{
-				MaxIdleConnsPerHost: conns,
-				MaxConnsPerHost:     conns,
-				IdleConnTimeout:     90 * time.Second,
-			},
-			Timeout: httpClientTimeout,
-		}
-		_httpSemaphore = make(chan struct{}, obidefault.ParallelWorkers())
-	})
-	return _httpClient
-}
-
-// RegisterHTTP registers the http module in the Lua state as a global,
-// consistent with obicontext and BioSequence.
-//
-// Exposes:
-//
-//	http.post(url, body [, timeout_ms]) → response string  (on success)
-//	http.post(url, body [, timeout_ms]) → nil, err string  (on error)
-//	http.set_concurrency(n)             → set max simultaneous requests
-func RegisterHTTP(luaState *lua.LState) {
-	table := luaState.NewTable()
-	luaState.SetField(table, "post", luaState.NewFunction(luaHTTPPost))
-	luaState.SetField(table, "set_concurrency", luaState.NewFunction(luaHTTPSetConcurrency))
-	luaState.SetGlobal("http", table)
-}
-
-// luaHTTPPost implements http.post(url, body [, timeout_ms]) for Lua.
-//
-// The optional third argument overrides the default timeout (in milliseconds).
-// Concurrent requests are throttled through _httpSemaphore so that a
-// single-threaded backend server is not overwhelmed by K parallel workers.
-//
-// Lua signature:
-//
-//	local response          = http.post(url, body)
-//	local response          = http.post(url, body, 5000)   -- 5 s timeout
-//	local response, err     = http.post(url, body)
-func luaHTTPPost(L *lua.LState) int {
-	url := L.CheckString(1)
-	body := L.CheckString(2)
-
-	client := getHTTPClient()
-
-	timeout := httpClientTimeout
-	if L.GetTop() >= 3 {
-		ms := L.CheckInt(3)
-		timeout = time.Duration(ms) * time.Millisecond
-	}
-
-	// Acquire semaphore slot — blocks until a slot is free.
-	_httpSemaphore <- struct{}{}
-	defer func() { <-_httpSemaphore }()
-
-	ctx, cancel := context.WithTimeout(context.Background(), timeout)
-	defer cancel()
-
-	req, err := http.NewRequestWithContext(ctx, http.MethodPost, url, strings.NewReader(body))
-	if err != nil {
-		L.Push(lua.LNil)
-		L.Push(lua.LString(err.Error()))
-		return 2
-	}
-	req.Header.Set("Content-Type", "application/json")
-
-	resp, err := client.Do(req)
-	if err != nil {
-		L.Push(lua.LNil)
-		L.Push(lua.LString(err.Error()))
-		return 2
-	}
-	defer resp.Body.Close()
-
-	respBytes, err := io.ReadAll(resp.Body)
-	if err != nil {
-		L.Push(lua.LNil)
-		L.Push(lua.LString(err.Error()))
-		return 2
-	}
-
-	L.Push(lua.LString(respBytes))
-	return 1
-}
-
-// luaHTTPSetConcurrency replaces the semaphore with a new one of size n.
-// Must be called before the first http.post (e.g. in begin()).
-//
-// Lua signature:
-//
-//	http.set_concurrency(1)   -- serialise all HTTP requests
-func luaHTTPSetConcurrency(L *lua.LState) int {
-	n := L.CheckInt(1)
-	if n < 1 {
-		n = 1
-	}
-	getHTTPClient() // ensure singleton is initialised
-	_httpSemaphore = make(chan struct{}, n)
-	return 0
-}

pkg/obilua/luajson.go

@@ -1,71 +0,0 @@
-package obilua
-
-import (
-	"encoding/json"
-
-	lua "github.com/yuin/gopher-lua"
-)
-
-// RegisterJSON registers the json module in the Lua state as a global,
-// consistent with obicontext, BioSequence, and http.
-//
-// Exposes:
-//
-//	json.encode(data)   → string         (on success)
-//	json.encode(data)   → nil, err       (on error)
-//	json.decode(string) → value          (on success)
-//	json.decode(string) → nil, err       (on error)
-func RegisterJSON(luaState *lua.LState) {
-	table := luaState.NewTable()
-	luaState.SetField(table, "encode", luaState.NewFunction(luaJSONEncode))
-	luaState.SetField(table, "decode", luaState.NewFunction(luaJSONDecode))
-	luaState.SetGlobal("json", table)
-}
-
-// luaJSONEncode implements json.encode(data) for Lua.
-func luaJSONEncode(L *lua.LState) int {
-	val := L.CheckAny(1)
-
-	var goVal interface{}
-	switch v := val.(type) {
-	case *lua.LTable:
-		goVal = Table2Interface(L, v)
-	case lua.LString:
-		goVal = string(v)
-	case lua.LNumber:
-		goVal = float64(v)
-	case lua.LBool:
-		goVal = bool(v)
-	case *lua.LNilType:
-		goVal = nil
-	default:
-		L.Push(lua.LNil)
-		L.Push(lua.LString("json.encode: unsupported type"))
-		return 2
-	}
-
-	b, err := json.Marshal(goVal)
-	if err != nil {
-		L.Push(lua.LNil)
-		L.Push(lua.LString(err.Error()))
-		return 2
-	}
-
-	L.Push(lua.LString(b))
-	return 1
-}
-
-// luaJSONDecode implements json.decode(string) for Lua.
-func luaJSONDecode(L *lua.LState) int {
-	s := L.CheckString(1)
-
-	var goVal interface{}
-	if err := json.Unmarshal([]byte(s), &goVal); err != nil {
-		L.Push(lua.LNil)
-		L.Push(lua.LString(err.Error()))
-		return 2
-	}
-
-	pushInterfaceToLua(L, goVal)
-	return 1
-}

pkg/obilua/luajson_test.go

@@ -1,184 +0,0 @@
-package obilua
-
-import (
-	"testing"
-
-	lua "github.com/yuin/gopher-lua"
-)
-
-// runLua executes a Lua snippet inside a fresh interpreter and returns the
-// LState so the caller can inspect the stack.
-func runLua(t *testing.T, script string) *lua.LState {
-	t.Helper()
-	L := NewInterpreter()
-	if err := L.DoString(script); err != nil {
-		t.Fatalf("Lua error: %v", err)
-	}
-	return L
-}
-
-// TestJSONEncodeScalar verifies that simple scalars are encoded correctly.
-func TestJSONEncodeScalar(t *testing.T) {
-	cases := []struct {
-		script   string
-		expected string
-	}{
-		{`result = json.encode("hello")`, `"hello"`},
-		{`result = json.encode(42)`, `42`},
-		{`result = json.encode(true)`, `true`},
-	}
-
-	for _, tc := range cases {
-		L := runLua(t, tc.script)
-		got := string(L.GetGlobal("result").(lua.LString))
-		if got != tc.expected {
-			t.Errorf("encode(%s): got %q, want %q", tc.script, got, tc.expected)
-		}
-		L.Close()
-	}
-}
-
-// TestJSONEncodeTable verifies that a Lua table (array and map) encodes to JSON.
-func TestJSONEncodeTable(t *testing.T) {
-	L := runLua(t, `result = json.encode({a = 1, b = "x"})`)
-	got := string(L.GetGlobal("result").(lua.LString))
-	// json.Marshal produces deterministic output for maps in Go 1.12+... actually not.
-	// Just check it round-trips via decode instead.
-	L.Close()
-	if got == "" {
-		t.Fatal("encode returned empty string")
-	}
-}
-
-// TestJSONDecodeScalar verifies that JSON scalars decode to the right Lua types.
-func TestJSONDecodeScalar(t *testing.T) {
-	L := runLua(t, `
-		s = json.decode('"hello"')
-		n = json.decode('3.14')
-		b = json.decode('true')
-	`)
-	if s, ok := L.GetGlobal("s").(lua.LString); !ok || string(s) != "hello" {
-		t.Errorf("decode string: got %v", L.GetGlobal("s"))
-	}
-	if n, ok := L.GetGlobal("n").(lua.LNumber); !ok || float64(n) != 3.14 {
-		t.Errorf("decode number: got %v", L.GetGlobal("n"))
-	}
-	if b, ok := L.GetGlobal("b").(lua.LBool); !ok || !bool(b) {
-		t.Errorf("decode bool: got %v", L.GetGlobal("b"))
-	}
-	L.Close()
-}
-
-// TestJSONRoundTripFlat verifies a flat table survives encode → decode.
-func TestJSONRoundTripFlat(t *testing.T) {
-	L := runLua(t, `
-		original = {name = "Homo_sapiens", score = 1.0, valid = true}
-		encoded  = json.encode(original)
-		decoded  = json.decode(encoded)
-	`)
-	decoded, ok := L.GetGlobal("decoded").(*lua.LTable)
-	if !ok {
-		t.Fatal("decoded is not a table")
-	}
-	if v := decoded.RawGetString("name"); string(v.(lua.LString)) != "Homo_sapiens" {
-		t.Errorf("name: got %v", v)
-	}
-	if v := decoded.RawGetString("score"); float64(v.(lua.LNumber)) != 1.0 {
-		t.Errorf("score: got %v", v)
-	}
-	if v := decoded.RawGetString("valid"); !bool(v.(lua.LBool)) {
-		t.Errorf("valid: got %v", v)
-	}
-	L.Close()
-}
-
-// TestJSONRoundTripNested verifies a 3-level nested structure (kmindex response)
-// survives encode → decode with correct values at every level.
-func TestJSONRoundTripNested(t *testing.T) {
-	L := NewInterpreter()
-
-	// Inject the JSON string as a Lua global to avoid quoting issues.
-	L.SetGlobal("kmindex_json", lua.LString(
-		`{"Human":{"query_001":{"Homo_sapiens--GCF_000001405_40":1.0}}}`,
-	))
-
-	if err := L.DoString(`
-		data      = json.decode(kmindex_json)
-		reencoded = json.encode(data)
-		data2     = json.decode(reencoded)
-	`); err != nil {
-		t.Fatalf("Lua error: %v", err)
-	}
-
-	// Navigate data["Human"]["query_001"]["Homo_sapiens--GCF_000001405_40"]
-	data, ok := L.GetGlobal("data").(*lua.LTable)
-	if !ok {
-		t.Fatal("data is not a table")
-	}
-	human, ok := data.RawGetString("Human").(*lua.LTable)
-	if !ok {
-		t.Fatal("data.Human is not a table")
-	}
-	query, ok := human.RawGetString("query_001").(*lua.LTable)
-	if !ok {
-		t.Fatal("data.Human.query_001 is not a table")
-	}
-	score, ok := query.RawGetString("Homo_sapiens--GCF_000001405_40").(lua.LNumber)
-	if !ok || float64(score) != 1.0 {
-		t.Errorf("score: got %v, want 1.0", query.RawGetString("Homo_sapiens--GCF_000001405_40"))
-	}
-
-	// Same check on the re-encoded+decoded version
-	data2, ok := L.GetGlobal("data2").(*lua.LTable)
-	if !ok {
-		t.Fatal("data2 is not a table")
-	}
-	score2 := data2.RawGetString("Human").(*lua.LTable).
-		RawGetString("query_001").(*lua.LTable).
-		RawGetString("Homo_sapiens--GCF_000001405_40").(lua.LNumber)
-	if float64(score2) != 1.0 {
-		t.Errorf("data2 score: got %v, want 1.0", score2)
-	}
-	L.Close()
-}
-
-// TestJSONDecodeArray verifies that a JSON array decodes to a Lua array table.
-func TestJSONDecodeArray(t *testing.T) {
-	L := runLua(t, `arr = json.decode('[1, 2, 3]')`)
-	arr, ok := L.GetGlobal("arr").(*lua.LTable)
-	if !ok {
-		t.Fatal("arr is not a table")
-	}
-	for i, expected := range []float64{1, 2, 3} {
-		v, ok := arr.RawGetInt(i + 1).(lua.LNumber)
-		if !ok || float64(v) != expected {
-			t.Errorf("arr[%d]: got %v, want %v", i+1, arr.RawGetInt(i+1), expected)
-		}
-	}
-	L.Close()
-}
-
-// TestJSONEncodeError verifies that json.encode on an unsupported type returns nil + error.
-func TestJSONEncodeError(t *testing.T) {
-	L := runLua(t, `
-		local result, err = json.encode(nil)
-	`)
-	// nil encodes to JSON "null" — not an error
-	L.Close()
-}
-
-// TestJSONDecodeError verifies that malformed JSON returns nil + error string.
-func TestJSONDecodeError(t *testing.T) {
-	L := runLua(t, `
-		local result, err = json.decode("not valid json")
-		decode_ok     = (result == nil)
-		decode_has_err = (err ~= nil)
-	`)
-	if L.GetGlobal("decode_ok") != lua.LTrue {
-		t.Error("expected nil result on decode error")
-	}
-	if L.GetGlobal("decode_has_err") != lua.LTrue {
-		t.Error("expected error string on decode error")
-	}
-	L.Close()
-}

pkg/obilua/obilib.go

@@ -5,6 +5,4 @@ import lua "github.com/yuin/gopher-lua"
 func RegisterObilib(luaState *lua.LState) {
 	RegisterObiSeq(luaState)
 	RegisterObiTaxonomy(luaState)
-	RegisterHTTP(luaState)
-	RegisterJSON(luaState)
 }

pkg/obilua/obiseqslice.go

@@ -31,8 +31,7 @@ func obiseqslice2Lua(interpreter *lua.LState,
 }
 
 func newObiSeqSlice(luaState *lua.LState) int {
-	capacity := luaState.OptInt(1, 0)
-	seqslice := obiseq.NewBioSequenceSlice(capacity)
+	seqslice := obiseq.NewBioSequenceSlice()
 	luaState.Push(obiseqslice2Lua(luaState, seqslice))
 	return 1
 }

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 4.4.41"
+var _Version = "Release 4.4.25"
 
 // Version returns the version of the obitools package.
 //

pkg/obiseq/biosequence.go

@@ -499,9 +499,6 @@ func (s *BioSequence) SetQualities(qualities Quality) {
 	if s.qualities != nil {
 		RecycleSlice(&s.qualities)
 	}
-	if len(qualities) > 0 && len(qualities) != len(s.sequence) {
-		log.Panicf("[BioSequence.SetQualities] Sequence %s has a length of %d and qualities a length of %d", s.id, len(s.sequence), len(qualities))
-	}
 	s.qualities = CopySlice(qualities)
 }
 
@@ -511,9 +508,6 @@ func (s *BioSequence) TakeQualities(qualities Quality) {
 	if s.qualities != nil {
 		RecycleSlice(&s.qualities)
 	}
-	if len(qualities) > 0 && len(qualities) != len(s.sequence) {
-		log.Panicf("[BioSequence.TakeQualities] Sequence %s has a length of %d and qualities a length of %d", s.id, len(s.sequence), len(qualities))
-	}
 	s.qualities = qualities
 }
 

pkg/obiseq/revcomp.go

@@ -118,9 +118,6 @@ func (sequence *BioSequence) _revcmpMutation() *BioSequence {
  */
 func ReverseComplementWorker(inplace bool) SeqWorker {
 	f := func(input *BioSequence) (BioSequenceSlice, error) {
-		if input.IsPaired() {
-			input.PairedWith().ReverseComplement(inplace)
-		}
 		return BioSequenceSlice{input.ReverseComplement(inplace)}, nil
 	}
 

pkg/obiseq/subseq.go

@@ -48,16 +48,7 @@ func (sequence *BioSequence) Subsequence(from, to int, circular bool) (*BioSeque
 		newSeq.sequence = CopySlice(sequence.Sequence()[from:to])
 
 		if sequence.HasQualities() {
-			qual := sequence.Qualities()
-			if len(qual) != sequence.Len() {
-				log.Panicf(
-					"[BioSequence.Subsequence] Sequence %s has a length of %d and qualities a length of %d",
-					sequence.Id(),
-					sequence.Len(),
-					len(qual),
-				)
-			}
-			newSeq.qualities = CopySlice(qual[from:to])
+			newSeq.qualities = CopySlice(sequence.Qualities()[from:to])
 		}
 
 		newSeq.id = fmt.Sprintf("%s_sub[%d..%d]", sequence.Id(), from+1, to)
@@ -67,16 +58,7 @@ func (sequence *BioSequence) Subsequence(from, to int, circular bool) (*BioSeque
 		newSeq.Write(sequence.Sequence()[0:to])
 
 		if sequence.HasQualities() {
-			qual := sequence.Qualities()
-			if len(qual) != sequence.Len() {
-				log.Panicf(
-					"[BioSequence.Subsequence] Sequence %s has a length of %d and qualities a length of %d",
-					sequence.Id(),
-					sequence.Len(),
-					len(qual),
-				)
-			}
-			newSeq.WriteQualities(qual[0:to])
+			newSeq.WriteQualities(sequence.Qualities()[0:to])
 		}
 
 	}

pkg/obiseq/worker.go

@@ -104,11 +104,11 @@ func SeqToSliceWorker(worker SeqWorker,
 			for _, s := range input {
 				r, err := worker(s)
 				if err == nil {
-					if i+len(r) > cap(output) {
-						output = slices.Grow(output[:i], len(r))
+					for _, rs := range r {
+						if i == len(output) {
+							output = slices.Grow(output, cap(output))
 							output = output[:cap(output)]
 						}
-					for _, rs := range r {
 						output[i] = rs
 						i++
 					}

pkg/obitools/obicsv/obicsv.go

@@ -33,7 +33,6 @@ func CLIWriteSequenceCSV(iterator obiiter.IBioSequence,
 		CSVSequence(CLIPrintSequence()),
 		CSVQuality(CLIPrintQuality()),
 		CSVAutoColumn(CLIAutoColumns()),
-		CSVNAValue(CLINAValue()),
 	)
 
 	csvIter := NewCSVSequenceIterator(iterator, opts...)

pkg/obitools/obicsv/sequence.go

@@ -1,7 +1,6 @@
 package obicsv
 
 import (
-	"fmt"
 	"log"
 	"slices"
 
@@ -68,19 +67,8 @@ func CSVBatchFromSequences(batch obiiter.BioSequenceBatch, opt Options) obiiterc
 
 			if taxon != nil {
 				taxid = taxon.String()
-			} else if ta, ok := sequence.GetAttribute("taxid"); ok {
-				switch tv := ta.(type) {
-				case string:
-					taxid = tv
-				case int:
-					taxid = fmt.Sprintf("%d", tv)
-				case float64:
-					taxid = fmt.Sprintf("%d", int(tv))
-				default:
-					taxid = opt.CSVNAValue()
-				}
 			} else {
-				taxid = opt.CSVNAValue()
+				taxid = sequence.Taxid()
 			}
 
 			record["taxid"] = taxid

pkg/obitools/obidistribute/distribute.go

@@ -46,7 +46,8 @@ func CLIDistributeSequence(sequences obiiter.IBioSequence) {
 		formater = obiformats.WriteSequencesToFile
 	}
 
-	dispatcher := sequences.Distribute(CLISequenceClassifier())
+	dispatcher := sequences.Distribute(CLISequenceClassifier(),
+		obidefault.BatchSize())
 
 	obiformats.WriterDispatcher(CLIFileNamePattern(),
 		dispatcher, formater, opts...,

pkg/obitools/obisummary/obisummary.go

@@ -99,17 +99,6 @@ func (data1 *DataSummary) Add(data2 *DataSummary) *DataSummary {
 	rep.sample_singletons = sumUpdateIntMap(data1.sample_singletons, data2.sample_singletons)
 	rep.sample_obiclean_bad = sumUpdateIntMap(data1.sample_obiclean_bad, data2.sample_obiclean_bad)
 
-	for k, m1 := range data1.map_summaries {
-		rep.map_summaries[k] = m1
-	}
-	for k, m2 := range data2.map_summaries {
-		if m1, ok := rep.map_summaries[k]; ok {
-			rep.map_summaries[k] = sumUpdateIntMap(m1, m2)
-		} else {
-			rep.map_summaries[k] = m2
-		}
-	}
-
 	return rep
 }
 
@@ -174,9 +163,8 @@ func ISummary(iterator obiiter.IBioSequence, summarise []string) map[string]inte
 	summaries := make([]*DataSummary, nproc)
 
 	for n := 0; n < nproc; n++ {
-		summaries[n] = NewDataSummary()
 		for _, v := range summarise {
-			summaries[n].map_summaries[v] = make(map[string]int)
+			summaries[n].map_summaries[v] = make(map[string]int, 0)
 		}
 	}
 
@@ -186,11 +174,6 @@ func ISummary(iterator obiiter.IBioSequence, summarise []string) map[string]inte
 			batch := iseq.Get()
 			for _, seq := range batch.Slice() {
 				summary.Update(seq)
-				for _, attr := range summarise {
-					if m, ok := seq.GetIntMap(attr); ok {
-						summary.map_summaries[attr] = sumUpdateIntMap(summary.map_summaries[attr], m)
-					}
-				}
 			}
 		}
 		waiter.Done()
@@ -198,9 +181,11 @@ func ISummary(iterator obiiter.IBioSequence, summarise []string) map[string]inte
 
 	waiter.Add(nproc)
 
+	summaries[0] = NewDataSummary()
 	go ff(iterator, summaries[0])
 
 	for i := 1; i < nproc; i++ {
+		summaries[i] = NewDataSummary()
 		go ff(iterator.Split(), summaries[i])
 	}
 
@@ -261,14 +246,5 @@ func ISummary(iterator obiiter.IBioSequence, summarise []string) map[string]inte
 			}
 		}
 	}
-
-	if len(rep.map_summaries) > 0 {
-		mapDict := make(map[string]interface{}, len(rep.map_summaries))
-		for attr, counts := range rep.map_summaries {
-			mapDict[attr] = counts
-		}
-		dict["map_summaries"] = mapDict
-	}
-
 	return dict
 }

prompt_documentation_globale.md

@@ -1,302 +0,0 @@
-# Objective
-
-Fully document OBITools (version 4, written in Go) in English, using a 4‑phase incremental pipeline.
-
-You **MUST** use the available MCP servers:
-
-- `cclsp` – exact definitions, references, diagnostics  
-- `jcodemunch` – code indexing, symbol extraction  
-- `treesitter` – AST and CLI parsing  
-- `context7` – external documentation  
-
-All tool calls must follow the exact API described in the MCP server documentation. If a required tool is unavailable, you **MUST** log the error and stop execution.
-
-### Tool call format (CRITICAL)
-
-Tool calls **MUST** use this exact XML format — no spaces inside the angle brackets:
-
-```
-<function=tool_name>
-{"param": "value"}
-</function>
-```
-
-**FORBIDDEN** — these variants will cause parse errors and must NEVER be used:
-- `< function=tool_name >` (spaces around the tag name)
-- `< function = tool_name>` (spaces around `=`)
-- `<function = tool_name>` (space before `=`)
-
-The opening tag is `<function=tool_name>` with **zero spaces** inside `<` and `>`.
-
----
-
-# Global rules
-
-** You are not allowed to read twice the same file in a row. **
-
-## Language
-
-- All generated documentation **MUST** be in English.  
-- If an existing documentation file is in French:  
-  1. Translate it to English  
-  2. Save the original as `.fr.md` **before** overwriting  
-  3. Write the new English version  
-
----
-
-## Execution mode (STRICT)
-
-You are operating in **STRICT TOOL MODE**:
-
-- If a file must be written, you **MUST** use the `Shell` tool.
-- You **MUST NOT** read entire directory listings into memory.
-- You **MUST** work with **one item at a time** using a simple text file as a task queue.
-
-### Reading files before writing
-
-- **Before writing to an existing documentation file**, you must first read it using the `Read` tool.
-- **When documenting a single Go source file**, you only need to read that one file (plus up to 4-5 helper files if needed for context).
-- Do NOT read the entire codebase - only what is necessary to document the current file.
-
----
-
-### Rules
-
-- Always write the **full** file (no partial updates).  
-- Paths are relative to the project root; directories are created implicitly.  
-- Content must be valid UTF‑8; use `\n` line endings.  
-- Do **not** wrap content in backticks.  
-
----
-
-## Progress tracking: task queue files
-
-We use **line‑oriented task files** to avoid loading large lists into memory. Each phase has its own task file:
-
-- `docs/todo/phase1.txt` – list of Go files (one per line) to document.  
-- `docs/todo/phase1bis.txt` – same list, but after phase1 is done.  
-- `docs/todo/phase2.txt` – list of packages.  
-- `docs/todo/phase3.txt` – list of tools.
-
-**How it works:**
-
-1. At the start of a phase, if the task file does not exist, it is created by scanning the codebase once (Phase 0 or Phase X init).  
-2. **Each run of the LLM processes only the first line of the task file.**  
-3. After processing the item (success or permanent failure), the line is removed from the task file.  
-   - On success, the line is deleted (no extra sentinel file needed).  
-   - On transient failure (retry < 3), we keep the line but increment a retry counter stored in a separate file.  
-   - On permanent failure (retry ≥ 3), we move the line to a `failed.txt` file and log the error.  
-4. The LLM then exits (or continues if the task file is still non‑empty, but it must never load more than one line).
-
-This way, the LLM’s context never holds more than a single task at a time.
-
-### Retry mechanism
-
-For each item (e.g., `internal/align/align.go`), we maintain a retry counter in:
-
-- `docs/retry/phase1/internal/align/align.go.count`
-
-If the file does not exist, retries = 0.  
-Each time processing fails, we increment the counter (write the new number).  
-If after increment the counter < 3, we keep the line in the task file.  
-If counter reaches 3, we **remove the line from the task file**, add it to `docs/failed/phase1/internal/align/align.go.failed` (just a marker), and log the error.
-
----
-
-## Documentation quality requirements (CRITICAL)
-
-Documentation MUST NOT be superficial. For each documented element (file, function, struct, package):
-
-###  You MUST explain:
-
-- what it does
-- why it exists (context, problem solved)
-- how it is used
-- assumptions and preconditions
-- possible edge cases
-
-### Forbidden patterns
-
-- Vague phrases like “This function handles…”, “Utility for…”, “Helper function…”.
-- Generic descriptions that could apply to any project.
-
-### Required content per element type
-
-- Functions:
-  - Purpose
-  - Parameter meaning
-  - Return values
-  - Notable behaviour (panic conditions, side effects, concurrency)
-- Structs:
-  - Role in the system
-  - Meaning of key fields
-- Files:
-  - Role within the package
-  - Interactions with other files
-
-### Anti‑generic rule
-
-If the description could apply to any project, it is INVALID. You MUST include domain‑specific context (bioinformatics, sequence processing, etc.) and concrete behaviour.
-
-### Quality validation
-
-Before marking an item as done (i.e., creating the .done sentinel), you MUST perform a self‑validation:
-
-- Check that all required sections are present.
-- Verify that no forbidden patterns remain.
-
-If validation fails, increment the retry counter and keep the item pending.
-
-
----
-
-# Directory structure
-
-```
-docs/
-  todo/                          # task queues
-    phase1.txt
-    phase1bis.txt
-    phase2.txt
-    phase3.txt
-  retry/                         # retry counters
-    phase1/                      # mirrors file structure
-      internal/align/align.go.count
-    phase1bis/
-    phase2/
-    phase3/
-  failed/                        # permanent failure markers
-    phase1/
-      internal/align/align.go.failed
-    phase1bis/
-    phase2/
-    phase3/
-  phase1/                        # actual documentation
-    <relative_path>/<file>.go.md
-  phase2/
-    <package>.md
-  phase3/
-    <tool>.md
-  error.log
-```
-
----
-
-# Phase 0: Initialization
-
-1. Ensure required directories exist: `docs/todo`, `docs/retry`, `docs/failed`, `docs/phase1`, `docs/phase2`, `docs/phase3`.  
-2. **If `docs/todo/phase1.txt` does not exist**:  
-   - Use `find pkg -name "*.go" ! -name "*_test.go" ! -path "*/cmd/*"` to list all Go files (excluding tests and main.go).  
-   - Write the list (one relative path per line, e.g., `internal/align/align.go`) to `docs/todo/phase1.txt`.  
-3. Do the same for phase2 and phase3 later when those phases start.  
-4. **No other state is stored.**
-
----
-
-# Phase 1: File documentation
-
-**Processing rule:**
-- Read the **first line** of `docs/todo/phase1.txt` (using `head -n 1`).  
-- If the file is empty, Phase 1 is complete → proceed to Phase 1bis initialization.  
-- Otherwise, process that single file.
-
-**Processing a file:**
-
-1. Let `relpath` be the line content (e.g., `internal/align/align.go`).  
-2. Check if a permanent failure marker exists at `docs/failed/phase1/${relpath}.failed`. If yes, remove the line from the task file and skip (line will be deleted). 
-3. If the documentation file `docs/phase1/${relpath}.go.md` exists go directly to its validation (step 6).
-4. Otherwise, generate documentation for that file (using MCP tools as before).  
-5. Write the documentation to `docs/phase1/${relpath}.go.md`.  
-6. Validate quality.  
-7. If validation succeeds:  
-   - Remove the line from the task file.  
-   - Remove any retry counter file for this item.  
-   - (No sentinel needed; the removal from todo indicates completion.)  
-8. If validation fails:  
-   - Increment retry counter:  
-     - If `docs/retry/phase1/${relpath}.count` does not exist, set to 1.  
-     - Else read it, add 1, write back.  
-   - If new counter >= 3:  
-     - Remove line from task file.  
-     - Create `docs/failed/phase1/${relpath}.failed`.  
-     - Log error.  
-   - If new counter < 3:  
-     - Keep the line in the task file (do nothing, it stays as first line for next run).  
-9. **Exit** (or stop if this was a single run). The next invocation will read the first line again (same if retry, or next if removed).
-
-**Important:**  
-- Do **not** read more than one line.  
-- Do **not** attempt to process multiple items in one run.  
-- The LLM should finish after handling one item.
-
----
-
-# Phase 1bis: Review and harmonization
-
-When Phase 1 is complete (i.e., `docs/todo/phase1.txt` empty), we initialize `docs/todo/phase1bis.txt` with the same list of files (the ones that succeeded).  
-But note: we need to know which files were successfully documented. Since we removed lines from `phase1.txt` on success, we need a record. The simplest is to reuse the same list but we can generate it by listing the existing `.go.md` files in `docs/phase1/` (since every successful file has a `.go.md`).  
-Thus, Phase 1bis initialization:
-
-- If `docs/todo/phase1bis.txt` does not exist, create it by listing all `.go.md` files under `docs/phase1/`, stripping the `docs/phase1/` prefix and the `.go.md` suffix, and writing the relative path (same format as phase1).  
-
-Then processing is identical to Phase 1, but using `docs/todo/phase1bis.txt` and output is overwriting the same `.go.md` files (with improvements). Retry counters go in `docs/retry/phase1bis/`.
-
----
-
-# Phase 2: Package documentation
-
-When Phase 1bis is complete (`docs/todo/phase1bis.txt` empty), initialize `docs/todo/phase2.txt`:
-
-- List all packages: unique directories under `pkg/` that contain at least one `.go` file and are not tools.  
-- Write each package identifier (e.g., `align`, `internal/align`) as a line.
-
-Processing: read first line, generate `docs/phase2/<package>.md`, validate, remove line on success, retry logic in `docs/retry/phase2/`.
-
----
-
-# Phase 3: Tool documentation
-
-When Phase 2 complete, initialize `docs/todo/phase3.txt`:
-
-- List all directories under `cmd/` that contain a `main.go`. Write each tool name as a line.
-
-Processing: read first line, generate `docs/phase3/<tool>.md`, validate, remove line on success, retry logic in `docs/retry/phase3/`.
-
----
-
-# Finalization
-
-When all task files are empty and no pending phases, generate `docs/README.md` by:
-
-- Listing all package docs (files in `docs/phase2/`) and linking.  
-- Listing all tool docs (files in `docs/phase3/`) and linking.  
-
-Write using `Shell`.
-
----
-
-# Execution flow summary
-
-1. **Phase 0**: Create directories and initial `todo/phase1.txt` if missing. Exit.  
-2. **Phase 1**:  
-   - If `todo/phase1.txt` exists and non‑empty → process first line.  
-   - Else → move to Phase 1bis initialization.  
-3. **Phase 1bis**:  
-   - If `todo/phase1bis.txt` does not exist → create from successful phase1 docs.  
-   - If non‑empty → process first line.  
-   - Else → move to Phase 2 initialization.  
-4. **Phase 2**: similar.  
-5. **Phase 3**: similar.  
-6. **Finalization**: generate README.
-
-The LLM should be invoked repeatedly (e.g., by a scheduler) until all phases are done. Each invocation processes exactly one item.
-
----
-
-# Important reminders
-
-- Always call `Shell` to write files; never output content in plain text.  
-- Validate quality before removing a line from the task file.  
-- Log all failures to `docs/error.log` in JSON lines format.  
-- If any MCP tool fails, treat as failure and increment retry counter.  
-- Never read more than one line from a task file in a single run.

scripts/find_setattribute.go

@@ -1,222 +0,0 @@
-//go:build ignore
-
-package main
-
-import (
-	"encoding/json"
-	"fmt"
-	"os"
-	"os/exec"
-	"path/filepath"
-	"regexp"
-	"sort"
-	"strconv"
-	"strings"
-)
-
-type Reference struct {
-	File     string `json:"file"`
-	Line     int    `json:"line"`
-	Column   int    `json:"column"`
-	Key      string `json:"key"`
-	Function string `json:"function"`
-	Context  string `json:"context"`
-}
-
-type Result struct {
-	Method     string      `json:"method"`
-	Signature  string      `json:"signature"`
-	Definition string      `json:"definition"`
-	References []Reference `json:"references"`
-	Total      int         `json:"total"`
-}
-
-var basePath = "/Users/coissac/Sync/travail/__MOI__/GO/obitools4"
-
-func main() {
-	cmd := exec.Command("rg", "-n", `\.SetAttribute\(`, basePath+"/pkg", "--type", "go")
-	output, err := cmd.Output()
-	if err != nil {
-		fmt.Fprintf(os.Stderr, "Error running rg: %v\n", err)
-		os.Exit(1)
-	}
-
-	lines := strings.Split(string(output), "\n")
-	lineRe := regexp.MustCompile(`^(.+?):(\d+):\s*(.+)$`)
-	keyRe := regexp.MustCompile(`SetAttribute\("([^"]+)"`)
-	templateKeyRe := regexp.MustCompile(`SetAttribute\("([^"]+)[^"]*"\s*,`)
-
-	var refs []Reference
-	seen := make(map[string]bool)
-
-	for _, line := range lines {
-		line = strings.TrimSpace(line)
-		if line == "" {
-			continue
-		}
-
-		matches := lineRe.FindStringSubmatch(line)
-		if matches == nil {
-			continue
-		}
-
-		file := matches[1]
-		lineNum, _ := strconv.Atoi(matches[2])
-		context := strings.TrimSpace(matches[3])
-
-		// Skip definition
-		if strings.Contains(file, "obiseq/attributes.go") && lineNum == 132 {
-			continue
-		}
-
-		// Extract key
-		var key string
-		if keyMatches := keyRe.FindStringSubmatch(context); keyMatches != nil {
-			key = keyMatches[1]
-		} else if tmplMatches := templateKeyRe.FindStringSubmatch(context); tmplMatches != nil {
-			key = tmplMatches[1]
-		} else {
-			continue
-		}
-
-		// Get function name using treesitter
-		funcName := getFunctionNameTreesitter(file, lineNum)
-
-		uniqueKey := fmt.Sprintf("%s:%d", file, lineNum)
-		if seen[uniqueKey] {
-			continue
-		}
-		seen[uniqueKey] = true
-
-		refs = append(refs, Reference{
-			File:     filepath.Base(file),
-			Line:     lineNum,
-			Column:   0,
-			Key:      key,
-			Function: funcName,
-			Context:  context,
-		})
-	}
-
-	sort.Slice(refs, func(i, j int) bool {
-		if refs[i].File != refs[j].File {
-			return refs[i].File < refs[j].File
-		}
-		return refs[i].Line < refs[j].Line
-	})
-
-	result := Result{
-		Method:     "SetAttribute",
-		Signature:  "func (s *BioSequence) SetAttribute(key string, value interface{})",
-		Definition: basePath + "/pkg/obiseq/attributes.go:132",
-		References: refs,
-		Total:      len(refs),
-	}
-
-	outputJSON, err := json.MarshalIndent(result, "", "  ")
-	if err != nil {
-		fmt.Fprintf(os.Stderr, "Error marshaling JSON: %v\n", err)
-		os.Exit(1)
-	}
-
-	fmt.Println(string(outputJSON))
-}
-
-// getFunctionNameTreesitter uses the treesitter_cursor_walk tool to get the containing function
-func getFunctionNameTreesitter(file string, targetLine int) string {
-	// Convert to 0-based for treesitter
-	row := targetLine - 1
-
-	// Use treesitter cursor walk to get ancestors
-	cmd := exec.Command("bash", "-c",
-		fmt.Sprintf(`kilo treesitter_cursor_walk --file_path %q --row %d --column 0 --max_depth 10 2>/dev/null`, file, row))
-
-	output, err := cmd.Output()
-	if err != nil {
-		return findContainingFunction(file, targetLine)
-	}
-
-	// Parse the JSON output to find function_declaration or method_declaration
-	var result map[string]interface{}
-	if err := json.Unmarshal(output, &result); err != nil {
-		return findContainingFunction(file, targetLine)
-	}
-
-	// Check ancestors for function declaration
-	if ancestors, ok := result["ancestors"].([]interface{}); ok {
-		for _, a := range ancestors {
-			if anc, ok := a.(map[string]interface{}); ok {
-				nodeType, _ := anc["type"].(string)
-				if nodeType == "function_declaration" || nodeType == "method_declaration" {
-					// Try to get the function name from children
-					if children, ok := anc["children"].([]interface{}); ok {
-						for _, c := range children {
-							if child, ok := c.(map[string]interface{}); ok {
-								childType, _ := child["type"].(string)
-								if childType == "identifier" {
-									if text, ok := child["text"].(string); ok {
-										return text
-									}
-								}
-								if childType == "field_identifier" {
-									if text, ok := child["text"].(string); ok {
-										return text
-									}
-								}
-							}
-						}
-					}
-				}
-				if nodeType == "func_literal" {
-					return "closure"
-				}
-			}
-		}
-	}
-
-	return findContainingFunction(file, targetLine)
-}
-
-func findContainingFunction(file string, targetLine int) string {
-	data, err := os.ReadFile(file)
-	if err != nil {
-		return ""
-	}
-	lines := strings.Split(string(data), "\n")
-
-	for i := targetLine - 1; i >= 0 && i >= targetLine-200; i-- {
-		if i >= len(lines) {
-			continue
-		}
-		line := strings.TrimSpace(lines[i])
-
-		if line == "}" && i > 0 {
-			for j := i - 1; j >= 0 && j >= i-50; j-- {
-				if j >= len(lines) {
-					continue
-				}
-				funcLine := strings.TrimSpace(lines[j])
-				if strings.HasPrefix(funcLine, "func ") {
-					if match := regexp.MustCompile(`func\s+\([^)]+\)\s+([a-zA-Z_][a-zA-Z0-9_]*)\s*\(`).FindStringSubmatch(funcLine); match != nil {
-						return match[1]
-					}
-					if match := regexp.MustCompile(`func\s+([a-zA-Z_][a-zA-Z0-9_]*)\s*\(`).FindStringSubmatch(funcLine); match != nil {
-						return match[1]
-					}
-				}
-			}
-			continue
-		}
-
-		if strings.HasPrefix(line, "func ") {
-			if match := regexp.MustCompile(`func\s+\([^)]+\)\s+([a-zA-Z_][a-zA-Z0-9_]*)\s*\(`).FindStringSubmatch(line); match != nil {
-				return match[1]
-			}
-			if match := regexp.MustCompile(`func\s+([a-zA-Z_][a-zA-Z0-9_]*)\s*\(`).FindStringSubmatch(line); match != nil {
-				return match[1]
-			}
-		}
-	}
-
-	return ""
-}

scripts/find_setattribute.sh

@@ -1,36 +0,0 @@
-#!/bin/bash
-
-basePath="/Users/coissac/Sync/travail/__MOI__/GO/obitools4"
-OUTPUT_FILE="${1:-/dev/stdout}"
-
-# Get all SetAttribute calls
-rg -n '\.SetAttribute\(' "$basePath/pkg" --type go | while read -r line; do
-	file="${line%%:*}"
-	line_num="${line%:*}"
-	line_num="${line_num##*:}"
-	context="${line##*: }"
-
-	# Extract key (only literal strings)
-	key=$(echo "$context" | sed -n 's/.*SetAttribute("\([^"]*\)".*/\1/p')
-	[ -z "$key" ] && continue
-
-	# Get function name using treesitter
-	func=$(kilo treesitter_cursor_walk \
-		--file_path "$file" \
-		--row "$((line_num - 1))" \
-		--column 0 \
-		--max_depth 10 2>/dev/null |
-		jq -r '.ancestors[] | select(.type == "function_declaration" or .type == "method_declaration") | .children[] | select(.type == "identifier" or .type == "field_identifier") | .text' 2>/dev/null)
-
-	# Fallback to func_literal for closures
-	if [ -z "$func" ]; then
-		func=$(kilo treesitter_cursor_walk \
-			--file_path "$file" \
-			--row "$((line_num - 1))" \
-			--column 0 \
-			--max_depth 10 2>/dev/null |
-			jq -r '.ancestors[] | select(.type == "func_literal") | "closure"' 2>/dev/null)
-	fi
-
-	echo "$(basename "$file")|$line_num|$key|${func:-unknown}|$context"
-done | sort -t'|' -k1,1 -k2,2n

setattribute_refs.json

@@ -1,308 +0,0 @@
-{
-  "obiannotate": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ],
-    "(git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiiter.IBioSequence).NumberSequences$1": [
-      "seq_number"
-    ]
-  },
-  "obiclean": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obicleandb": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetCount": [
-      "count"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obicomplement": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obiconsensus": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetCount": [
-      "count"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ],
-    "git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obitools/obiconsensus.BuildConsensus": [
-      "obiconsensus_kmer_max_occur",
-      "obiconsensus_filtered_graph_size",
-      "obiconsensus_full_graph_size",
-      "obiconsensus_consensus",
-      "obiconsensus_weight",
-      "obiconsensus_seq_length",
-      "obiconsensus_kmer_size"
-    ],
-    "git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obitools/obiconsensus.MinionClusterDenoise": [
-      "obiconsensus_consensus"
-    ],
-    "git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obitools/obiconsensus.MinionDenoise$1": [
-      "obiconsensus_consensus",
-      "obiconsensus_weight"
-    ]
-  },
-  "obiconvert": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obicount": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obicsv": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obidemerge": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obidistribute": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obigrep": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obijoin": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obikmermatch": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obikmersimcount": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obilandmark": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetCoordinate": [
-      "landmark_coord"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetOBITagGeomRefIndex": [
-      "obitag_geomref_index"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ],
-    "git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obitools/obilandmark.CLISelectLandmarkSequences": [
-      "landmark_id"
-    ]
-  },
-  "obimatrix": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obimicrosat": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obimultiplex": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obipairing": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence)._revcmpMutation": [
-      "pairing_mismatches"
-    ],
-    "git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obialign.BuildQualityConsensus": [
-      "pairing_mismatches"
-    ]
-  },
-  "obipcr": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obireffamidx": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetOBITagRefIndex": [
-      "obitag_ref_index"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ],
-    "git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obitools/obirefidx.IndexFamilyDB": [
-      "reffamidx_id"
-    ]
-  },
-  "obirefidx": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetOBITagRefIndex": [
-      "obitag_ref_index"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obiscript": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obisplit": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obisummary": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obitag": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetPath": [
-      "taxonomic_path"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  },
-  "obitagpcr": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obingslibrary.NGSLibrary).ExtractMultiBarcode": [
-      "obimultiplex_error",
-      "obimultiplex_amplicon_rank"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence)._revcmpMutation": [
-      "pairing_mismatches"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence)._subseqMutation": [
-      "pairing_mismatches"
-    ],
-    "git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obialign.BuildQualityConsensus": [
-      "pairing_mismatches"
-    ]
-  },
-  "obitaxonomy": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetPath": [
-      "taxonomic_path"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ],
-    "(git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiiter.IBioSequence).NumberSequences$1": [
-      "seq_number"
-    ]
-  },
-  "obiuniq": {
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetCount": [
-      "count"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetDefinition": [
-      "definition"
-    ],
-    "(*git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq.BioSequence).SetTaxid": [
-      "taxid"
-    ]
-  }
-}

version.txt

@@ -1 +1 @@
-4.4.41
+4.4.25

x

@@ -1,19 +0,0 @@
-```markdown
-# DNA Scoring and Matching Utilities in `obialign`
-
-This module provides low-level utilities for computing nucleotide alignment scores using probabilistic and bit-encoded representations.
-
-- **Bit Encoding**: Nucleotides are encoded in 4-bit groups (e.g., `A=0b0001`, `C=0b0010`, etc.), enabling efficient bitwise comparison.
-- **`_MatchRatio(a, b)`**: Computes a normalized match ratio between two encoded bytes based on shared bits:
-  `ratio = common_bits / (bits_in_a × bits_in_b)`.
-- **`_FourBitsCount`**: Precomputed lookup table for Hamming weight (popcount) of 4-bit values.
-- **Log-space Arithmetic**: Helper functions (`_Logaddexp`, `_Logdiffexp`, `_Log1mexp`) ensure numerical stability in probabilistic computations.
-- **Phred-scaled Quality Integration**:
-  `_MatchScoreRatio(QF, QR)` derives log-odds match/mismatch scores from Phred quality values (`QF`, `QR`), modeling sequencing error probabilities.
-- **Precomputed Matrices**:
-  - `_NucPartMatch[i][j]`: Match ratios for all nucleotide pairs (from 4-bit codes).
-  - `_NucScorePartMatchMatch/Mismatch[i][j]`: Integer-scaled match/mismatch scores (×10) for quality pairs `(i, j)` in `[0..99]`.
-- **Thread-Safe Initialization**: `_InitDNAScoreMatrix()` ensures one-time, synchronized initialization of all scoring tables via a mutex.
-
-Designed for high-performance alignment kernels where speed and numerical robustness are critical.
-```