fetch_ml/internal/api/protocol.go
Jeremie Fraeys 2e701340e5
feat(core): API, worker, queue, and manifest improvements
- Add protocol buffer optimizations (internal/api/protocol.go)
- Add filesystem queue backend (internal/queue/filesystem_queue.go)
- Add run manifest support (internal/manifest/run_manifest.go)
- Worker and jupyter task refinements
- Exported test wrappers for benchmarking
2026-02-12 12:05:17 -05:00

358 lines
8.8 KiB
Go

package api
import (
"encoding/binary"
"encoding/json"
"fmt"
"sync"
"time"
)
var bufferPool = sync.Pool{
New: func() interface{} {
buf := make([]byte, 0, 256)
return &buf
},
}
// Response packet types
const (
PacketTypeSuccess = 0x00
PacketTypeError = 0x01
PacketTypeProgress = 0x02
PacketTypeStatus = 0x03
PacketTypeData = 0x04
PacketTypeLog = 0x05
)
// Error codes
const (
ErrorCodeUnknownError = 0x00
ErrorCodeInvalidRequest = 0x01
ErrorCodeAuthenticationFailed = 0x02
ErrorCodePermissionDenied = 0x03
ErrorCodeResourceNotFound = 0x04
ErrorCodeResourceAlreadyExists = 0x05
ErrorCodeServerOverloaded = 0x10
ErrorCodeDatabaseError = 0x11
ErrorCodeNetworkError = 0x12
ErrorCodeStorageError = 0x13
ErrorCodeTimeout = 0x14
ErrorCodeJobNotFound = 0x20
ErrorCodeJobAlreadyRunning = 0x21
ErrorCodeJobFailedToStart = 0x22
ErrorCodeJobExecutionFailed = 0x23
ErrorCodeJobCancelled = 0x24
ErrorCodeOutOfMemory = 0x30
ErrorCodeDiskFull = 0x31
ErrorCodeInvalidConfiguration = 0x32
ErrorCodeServiceUnavailable = 0x33
)
// Progress types
const (
ProgressTypePercentage = 0x00
ProgressTypeStage = 0x01
ProgressTypeMessage = 0x02
ProgressTypeBytesTransferred = 0x03
)
// Log levels
const (
LogLevelDebug = 0x00
LogLevelInfo = 0x01
LogLevelWarn = 0x02
LogLevelError = 0x03
)
// ResponsePacket represents a structured response packet
type ResponsePacket struct {
PacketType byte
Timestamp uint64
// Success fields
SuccessMessage string
// Error fields
ErrorCode byte
ErrorMessage string
ErrorDetails string
// Progress fields
ProgressType byte
ProgressValue uint32
ProgressTotal uint32
ProgressMessage string
// Status fields
StatusData string
// Data fields
DataType string
DataPayload []byte
// Log fields
LogLevel byte
LogMessage string
}
// NewSuccessPacket creates a success response packet
func NewSuccessPacket(message string) *ResponsePacket {
return &ResponsePacket{
PacketType: PacketTypeSuccess,
Timestamp: uint64(time.Now().Unix()),
SuccessMessage: message,
}
}
// NewSuccessPacketWithPayload creates a success response packet with JSON payload
func NewSuccessPacketWithPayload(message string, payload interface{}) *ResponsePacket {
// Convert payload to JSON for the DataPayload field
payloadBytes, _ := json.Marshal(payload)
return &ResponsePacket{
PacketType: PacketTypeData,
Timestamp: uint64(time.Now().Unix()),
SuccessMessage: message,
DataType: "status",
DataPayload: payloadBytes,
}
}
// NewErrorPacket creates an error response packet
func NewErrorPacket(errorCode byte, message string, details string) *ResponsePacket {
return &ResponsePacket{
PacketType: PacketTypeError,
Timestamp: uint64(time.Now().Unix()),
ErrorCode: errorCode,
ErrorMessage: message,
ErrorDetails: details,
}
}
// NewProgressPacket creates a progress response packet
func NewProgressPacket(
progressType byte,
value uint32,
total uint32,
message string,
) *ResponsePacket {
return &ResponsePacket{
PacketType: PacketTypeProgress,
Timestamp: uint64(time.Now().Unix()),
ProgressType: progressType,
ProgressValue: value,
ProgressTotal: total,
ProgressMessage: message,
}
}
// NewStatusPacket creates a status response packet
func NewStatusPacket(data string) *ResponsePacket {
return &ResponsePacket{
PacketType: PacketTypeStatus,
Timestamp: uint64(time.Now().Unix()),
StatusData: data,
}
}
// NewDataPacket creates a data response packet
func NewDataPacket(dataType string, payload []byte) *ResponsePacket {
return &ResponsePacket{
PacketType: PacketTypeData,
Timestamp: uint64(time.Now().Unix()),
DataType: dataType,
DataPayload: payload,
}
}
// NewLogPacket creates a log response packet
func NewLogPacket(level byte, message string) *ResponsePacket {
return &ResponsePacket{
PacketType: PacketTypeLog,
Timestamp: uint64(time.Now().Unix()),
LogLevel: level,
LogMessage: message,
}
}
// Serialize converts the packet to binary format
func (p *ResponsePacket) Serialize() ([]byte, error) {
// For small packets, avoid pool overhead
if p.estimatedSize() <= 1024 {
buf := make([]byte, 0, p.estimatedSize())
return serializePacketToBuffer(p, buf)
}
// Use pool for larger packets
bufPtr := bufferPool.Get().(*[]byte)
defer func() {
*bufPtr = (*bufPtr)[:0]
bufferPool.Put(bufPtr)
}()
buf := *bufPtr
// Ensure buffer has enough capacity
if cap(buf) < p.estimatedSize() {
buf = make([]byte, 0, p.estimatedSize())
} else {
buf = buf[:0]
}
return serializePacketToBuffer(p, buf)
}
func serializePacketToBuffer(p *ResponsePacket, buf []byte) ([]byte, error) {
// Packet type
buf = append(buf, p.PacketType)
// Timestamp (8 bytes, big-endian)
var timestampBytes [8]byte
binary.BigEndian.PutUint64(timestampBytes[:], p.Timestamp)
buf = append(buf, timestampBytes[:]...)
// Packet-specific data
switch p.PacketType {
case PacketTypeSuccess:
buf = appendString(buf, p.SuccessMessage)
case PacketTypeError:
buf = append(buf, p.ErrorCode)
buf = appendString(buf, p.ErrorMessage)
buf = appendString(buf, p.ErrorDetails)
case PacketTypeProgress:
buf = append(buf, p.ProgressType)
buf = appendUint32(buf, p.ProgressValue)
buf = appendUint32(buf, p.ProgressTotal)
buf = appendString(buf, p.ProgressMessage)
case PacketTypeStatus:
buf = appendString(buf, p.StatusData)
case PacketTypeData:
buf = appendString(buf, p.DataType)
buf = appendBytes(buf, p.DataPayload)
case PacketTypeLog:
buf = append(buf, p.LogLevel)
buf = appendString(buf, p.LogMessage)
default:
return nil, fmt.Errorf("unknown packet type: %d", p.PacketType)
}
return buf, nil
}
// appendString writes a string with varint length prefix
func appendString(buf []byte, s string) []byte {
length := uint64(len(s))
var tmp [binary.MaxVarintLen64]byte
n := binary.PutUvarint(tmp[:], length)
buf = append(buf, tmp[:n]...)
return append(buf, s...)
}
// appendBytes writes bytes with varint length prefix
func appendBytes(buf []byte, b []byte) []byte {
length := uint64(len(b))
var tmp [binary.MaxVarintLen64]byte
n := binary.PutUvarint(tmp[:], length)
buf = append(buf, tmp[:n]...)
return append(buf, b...)
}
func appendUint32(buf []byte, value uint32) []byte {
var tmp [4]byte
binary.BigEndian.PutUint32(tmp[:], value)
return append(buf, tmp[:]...)
}
func (p *ResponsePacket) estimatedSize() int {
base := 1 + 8 // packet type + timestamp
switch p.PacketType {
case PacketTypeSuccess:
return base + binary.MaxVarintLen64 + len(p.SuccessMessage)
case PacketTypeError:
return base + 1 + 2*binary.MaxVarintLen64 + len(p.ErrorMessage) + len(p.ErrorDetails)
case PacketTypeProgress:
return base + 1 + 4 + 4 + binary.MaxVarintLen64 + len(p.ProgressMessage)
case PacketTypeStatus:
return base + binary.MaxVarintLen64 + len(p.StatusData)
case PacketTypeData:
return base + binary.MaxVarintLen64 + len(p.DataType) + binary.MaxVarintLen64 + len(p.DataPayload)
case PacketTypeLog:
return base + 1 + binary.MaxVarintLen64 + len(p.LogMessage)
default:
return base
}
}
// GetErrorMessage returns a human-readable error message for an error code
func GetErrorMessage(code byte) string {
switch code {
case ErrorCodeUnknownError:
return "Unknown error occurred"
case ErrorCodeInvalidRequest:
return "Invalid request format"
case ErrorCodeAuthenticationFailed:
return "Authentication failed"
case ErrorCodePermissionDenied:
return "Permission denied"
case ErrorCodeResourceNotFound:
return "Resource not found"
case ErrorCodeResourceAlreadyExists:
return "Resource already exists"
case ErrorCodeServerOverloaded:
return "Server is overloaded"
case ErrorCodeDatabaseError:
return "Database error occurred"
case ErrorCodeNetworkError:
return "Network error occurred"
case ErrorCodeStorageError:
return "Storage error occurred"
case ErrorCodeTimeout:
return "Operation timed out"
case ErrorCodeJobNotFound:
return "Job not found"
case ErrorCodeJobAlreadyRunning:
return "Job is already running"
case ErrorCodeJobFailedToStart:
return "Job failed to start"
case ErrorCodeJobExecutionFailed:
return "Job execution failed"
case ErrorCodeJobCancelled:
return "Job was cancelled"
case ErrorCodeOutOfMemory:
return "Server out of memory"
case ErrorCodeDiskFull:
return "Server disk full"
case ErrorCodeInvalidConfiguration:
return "Invalid server configuration"
case ErrorCodeServiceUnavailable:
return "Service temporarily unavailable"
default:
return "Unknown error code"
}
}
// GetLogLevelName returns the name for a log level
func GetLogLevelName(level byte) string {
switch level {
case LogLevelDebug:
return "DEBUG"
case LogLevelInfo:
return "INFO"
case LogLevelWarn:
return "WARN"
case LogLevelError:
return "ERROR"
default:
return "UNKNOWN"
}
}