Add non-working partial yzma implementation

go.mod

@@ -1,3 +1,11 @@
 module git.chandlerswift.com/chandlerswift/svs-services-server
 
 go 1.25.4
+
+require github.com/hybridgroup/yzma v1.3.0
+
+require (
+	github.com/ebitengine/purego v0.9.1 // indirect
+	github.com/jupiterrider/ffi v0.5.1 // indirect
+	golang.org/x/sys v0.38.0 // indirect
+)

go.sum

@@ -0,0 +1,8 @@
+github.com/ebitengine/purego v0.9.1 h1:a/k2f2HQU3Pi399RPW1MOaZyhKJL9w/xFpKAg4q1s0A=
+github.com/ebitengine/purego v0.9.1/go.mod h1:iIjxzd6CiRiOG0UyXP+V1+jWqUXVjPKLAI0mRfJZTmQ=
+github.com/hybridgroup/yzma v1.3.0 h1:5dw9qEcFEGEJq+tA12Ooa6D/e0PROqv7Ix6VfSR9MQI=
+github.com/hybridgroup/yzma v1.3.0/go.mod h1:UUYw+DLlrgtBYm+B+9XD3boB1ZcDpfbAnYHKW3VKKZ4=
+github.com/jupiterrider/ffi v0.5.1 h1:l7ANXU+Ex33LilVa283HNaf/sTzCrrht7D05k6T6nlc=
+github.com/jupiterrider/ffi v0.5.1/go.mod h1:x7xdNKo8h0AmLuXfswDUBxUsd2OqUP4ekC8sCnsmbvo=
+golang.org/x/sys v0.38.0 h1:3yZWxaJjBmCWXqhN1qh02AkOnCQ1poK6oF+a7xWL6Gc=
+golang.org/x/sys v0.38.0/go.mod h1:OgkHotnGiDImocRcuBABYBEXf8A9a87e/uXjp9XT3ks=

main.go

@@ -2,9 +2,12 @@ package main
 
 import (
 	"embed"
+	"fmt"
 	"html/template"
 	"net/http"
 	"time"
+
+	"github.com/hybridgroup/yzma/pkg/llama"
 )
 
 //go:embed templates/index.html
@@ -99,5 +102,47 @@ func main() {
 	// Serve embedded headshots file
 	http.Handle("/headshots/", http.FileServer(http.FS(headshots)))
 
+	// TODO: derive from something?
+	libPath := "/nix/store/jml3vhvay9yy94qj8bmmhbf2dhx6q2n1-llama-cpp-7356/lib"
+	modelFile := "./SmolLM-135M.Q2_K.gguf"
+	prompt := "Are you ready to go?"
+	responseLength := int32(128)
+
+	llama.Load(libPath)
+	llama.LogSet(llama.LogSilent())
+	llama.Init()
+
+	model, _ := llama.ModelLoadFromFile(modelFile, llama.ModelDefaultParams())
+	lctx, _ := llama.InitFromModel(model, llama.ContextDefaultParams())
+
+	vocab := llama.ModelGetVocab(model)
+
+	// get tokens from the prompt
+	tokens := llama.Tokenize(vocab, prompt, true, false)
+
+	batch := llama.BatchGetOne(tokens)
+
+	sampler := llama.SamplerChainInit(llama.SamplerChainDefaultParams())
+	llama.SamplerChainAdd(sampler, llama.SamplerInitGreedy())
+
+	for pos := int32(0); pos < responseLength; pos += batch.NTokens {
+		llama.Decode(lctx, batch)
+		token := llama.SamplerSample(sampler, lctx, -1)
+
+		if llama.VocabIsEOG(vocab, token) {
+			fmt.Println()
+			break
+		}
+
+		buf := make([]byte, 36)
+		len := llama.TokenToPiece(vocab, token, buf, 0, true)
+
+		fmt.Print(string(buf[:len]))
+
+		batch = llama.BatchGetOne([]llama.Token{token})
+	}
+
+	fmt.Println()
+
 	http.ListenAndServe(":8080", nil)
 }