feat: full BaraQL (GROUP BY/JOIN/CTE/aggregates), Raft consensus, sharding — 107 tests

BaraQL parser extended:
- GROUP BY + HAVING clauses
- JOIN (INNER, LEFT, RIGHT, FULL, CROSS) with ON conditions
- CTE (WITH ... AS) with multiple CTEs
- Aggregate functions (count, sum, avg, min, max)
- CASE/WHEN/THEN/ELSE/END expressions
- BETWEEN, IN, LIKE/ILIKE, IS NULL/IS NOT NULL
- Dotted path identifiers (table.column)
- Subqueries in FROM clause
- CREATE TYPE with colon syntax (name: type)
- UPDATE SET ... WHERE, DELETE FROM ... WHERE

Infrastructure:
- Raft consensus: leader election, log replication, RequestVote/AppendEntries
- Sharding: hash-based, range-based, consistent hashing, rebalancing
- Fixed lexer: = is equality, := is assignment, added : (colon) token
- 34 new parser tests (107 total, all passing)

ROADMAP.md

@@ -78,12 +78,12 @@
 - [x] Array литерали
 - [x] Типов анализатор (type checker)
 - [x] IR (Intermediate Representation)
-- [ ] Оптимизатор на заявки (predicate pushdown, projection pushdown)
+- [x] Оптимизатор на заявки (predicate pushdown, projection pushdown)
+- [x] GROUP BY, HAVING
+- [x] JOIN (inner, left, right, full, cross)
+- [x] CTE (WITH)
+- [x] Агрегатни функции (count, sum, avg, min, max)
 - [ ] Codegen → storage операции
-- [ ] GROUP BY, HAVING
-- [ ] JOIN (inner, left, right, full)
-- [ ] CTE (WITH)
-- [ ] Агрегатни функции (count, sum, avg, min, max)
 - [ ] Потребителски функции (UDF)
 
 ### Фаза 3: Мултимодален storage 🟡
@@ -170,11 +170,11 @@
 - [ ] Interactive query editor с autocomplete
 - [ ] Import/Export (JSON, CSV, Parquet)
 
-### Фаза 11: Кластеризация и разпределение ⬜
-- [ ] Raft консенсус протокол
-- [ ] Sharding (hash-based, range-based)
+### Фаза 11: Кластеризация и разпределение 🟡
+- [x] Raft консенсус протокол (leader election + log replication)
+- [x] Sharding (hash-based, range-based, consistent hashing)
 - [ ] Replication (sync, async)
-- [ ] Leader election
+- [ ] Leader election за multi-node
 - [ ] Gossip protocol за membership
 - [ ] Distributed transactions
 - [ ] Auto-rebalancing
@@ -196,16 +196,16 @@
 | Фаза | Статус | Напредък |
 |------|--------|----------|
 | 1. Ядро | ✅ Завършена | 95% |
-| 2. BaraQL | 🟡 В процес | 60% |
+| 2. BaraQL | ✅ Основно завършена | 85% |
 | 3. Мултимодален storage | 🟡 В процес | 75% |
 | 4. Транзакции | ✅ Основно завършена | 85% |
-| 5. Протокол | 🟡 В процес | 85% |
+| 5. Протокол | ✅ Основно завършена | 85% |
 | 6. Schema | ✅ Основно завършена | 75% |
 | 7. Векторен engine | ✅ Завършена | 95% |
 | 8. Graph engine | ✅ Завършена | 90% |
 | 9. FTS | ✅ Завършена | 85% |
 | 10. Клиенти и CLI | 🟡 В процес | 50% |
-| 11. Кластер | ⬜ Не стартирана | 0% |
+| 11. Кластер | 🟡 В процес | 30% |
 | 12. Оптимизации | ⬜ Не стартирана | 0% |
 
 **Легенда:** ⬜ Не стартирана | 🟡 В процес | ✅ Завършена

src/barabadb/core/raft.nim

@@ -0,0 +1,280 @@
+## Raft Consensus — leader election + log replication
+import std/tables
+import std/sets
+import std/deques
+import std/algorithm
+import std/random
+
+type
+  RaftState* = enum
+    rsFollower
+    rsCandidate
+    rsLeader
+
+  LogEntry* = object
+    term*: uint64
+    index*: uint64
+    command*: string
+    data*: seq[byte]
+
+  RaftNode* = ref object
+    id*: string
+    state*: RaftState
+    currentTerm*: uint64
+    votedFor*: string
+    log*: seq[LogEntry]
+    commitIndex*: uint64
+    lastApplied*: uint64
+    # Leader state
+    nextIndex*: Table[string, uint64]
+    matchIndex*: Table[string, uint64]
+    # Cluster
+    peers*: seq[string]
+    leaderId*: string
+    # Timing
+    electionTimeout*: int
+    heartbeatTimeout*: int
+    votesReceived*: HashSet[string]
+
+  RaftMessageKind* = enum
+    rmkRequestVote
+    rmkRequestVoteReply
+    rmkAppendEntries
+    rmkAppendEntriesReply
+
+  RaftMessage* = object
+    kind*: RaftMessageKind
+    term*: uint64
+    senderId*: string
+    # RequestVote
+    lastLogIndex*: uint64
+    lastLogTerm*: uint64
+    # AppendEntries
+    prevLogIndex*: uint64
+    prevLogTerm*: uint64
+    entries*: seq[LogEntry]
+    leaderCommit*: uint64
+    # Reply
+    success*: bool
+    matchIdx*: uint64
+
+  RaftCluster* = ref object
+    nodes*: Table[string, RaftNode]
+    messageQueue*: Deque[RaftMessage]
+
+proc newRaftNode*(id: string, peers: seq[string]): RaftNode =
+  RaftNode(
+    id: id,
+    state: rsFollower,
+    currentTerm: 0,
+    votedFor: "",
+    log: @[],
+    commitIndex: 0,
+    lastApplied: 0,
+    nextIndex: initTable[string, uint64](),
+    matchIndex: initTable[string, uint64](),
+    peers: peers,
+    leaderId: "",
+    electionTimeout: 150 + rand(150),
+    heartbeatTimeout: 50,
+    votesReceived: initHashSet[string](),
+  )
+
+proc newRaftCluster*(): RaftCluster =
+  RaftCluster(
+    nodes: initTable[string, RaftNode](),
+    messageQueue: initDeque[RaftMessage](),
+  )
+
+proc addNode*(cluster: RaftCluster, id: string) =
+  var peers: seq[string] = @[]
+  for existingId in cluster.nodes.keys:
+    peers.add(existingId)
+    cluster.nodes[existingId].peers.add(id)
+  cluster.nodes[id] = newRaftNode(id, peers)
+
+proc lastLogIndex*(node: RaftNode): uint64 =
+  if node.log.len == 0:
+    return 0
+  return node.log[^1].index
+
+proc lastLogTerm*(node: RaftNode): uint64 =
+  if node.log.len == 0:
+    return 0
+  return node.log[^1].term
+
+proc becomeFollower*(node: RaftNode, term: uint64) =
+  node.state = rsFollower
+  node.currentTerm = term
+  node.votedFor = ""
+  node.votesReceived.clear()
+
+proc becomeCandidate*(node: RaftNode) =
+  node.state = rsCandidate
+  inc node.currentTerm
+  node.votedFor = node.id
+  node.votesReceived.clear()
+  node.votesReceived.incl(node.id)
+
+proc becomeLeader*(node: RaftNode) =
+  node.state = rsLeader
+  node.leaderId = node.id
+  for peer in node.peers:
+    node.nextIndex[peer] = node.lastLogIndex + 1
+    node.matchIndex[peer] = 0
+
+proc handleRequestVote*(node: RaftNode, msg: RaftMessage): RaftMessage =
+  var reply = RaftMessage(
+    kind: rmkRequestVoteReply,
+    term: node.currentTerm,
+    senderId: node.id,
+    success: false,
+  )
+
+  if msg.term < node.currentTerm:
+    return reply
+
+  if msg.term > node.currentTerm:
+    node.becomeFollower(msg.term)
+
+  let canVote = node.votedFor == "" or node.votedFor == msg.senderId
+  let logOk = msg.lastLogTerm > node.lastLogTerm or
+              (msg.lastLogTerm == node.lastLogTerm and msg.lastLogIndex >= node.lastLogIndex)
+
+  if canVote and logOk:
+    node.votedFor = msg.senderId
+    reply.success = true
+    reply.term = node.currentTerm
+
+  return reply
+
+proc handleAppendEntries*(node: RaftNode, msg: RaftMessage): RaftMessage =
+  var reply = RaftMessage(
+    kind: rmkAppendEntriesReply,
+    term: node.currentTerm,
+    senderId: node.id,
+    success: false,
+    matchIdx: 0,
+  )
+
+  if msg.term < node.currentTerm:
+    return reply
+
+  if msg.term >= node.currentTerm:
+    node.becomeFollower(msg.term)
+    node.leaderId = msg.senderId
+
+  # Check if log contains entry at prevLogIndex with prevLogTerm
+  if msg.prevLogIndex > 0:
+    if msg.prevLogIndex > uint64(node.log.len):
+      return reply
+    if node.log[msg.prevLogIndex - 1].term != msg.prevLogTerm:
+      # Delete conflicting entries
+      node.log.setLen(int(msg.prevLogIndex - 1))
+      return reply
+
+  # Append new entries
+  for entry in msg.entries:
+    let idx = int(entry.index - 1)
+    if idx < node.log.len:
+      if node.log[idx].term != entry.term:
+        node.log.setLen(idx)
+        node.log.add(entry)
+    else:
+      node.log.add(entry)
+
+  # Update commit index
+  if msg.leaderCommit > node.commitIndex:
+    node.commitIndex = min(msg.leaderCommit, node.lastLogIndex)
+
+  reply.success = true
+  reply.matchIdx = node.lastLogIndex
+  return reply
+
+proc requestVote*(node: RaftNode): seq[RaftMessage] =
+  result = @[]
+  for peer in node.peers:
+    result.add(RaftMessage(
+      kind: rmkRequestVote,
+      term: node.currentTerm,
+      senderId: node.id,
+      lastLogIndex: node.lastLogIndex,
+      lastLogTerm: node.lastLogTerm,
+    ))
+
+proc appendEntries*(node: RaftNode, peerId: string): RaftMessage =
+  let nextIdx = node.nextIndex.getOrDefault(peerId, node.lastLogIndex + 1)
+  let prevIdx = nextIdx - 1
+  var prevTerm: uint64 = 0
+  if prevIdx > 0 and prevIdx <= uint64(node.log.len):
+    prevTerm = node.log[prevIdx - 1].term
+
+  var entries: seq[LogEntry] = @[]
+  for i in int(nextIdx - 1)..<node.log.len:
+    entries.add(node.log[i])
+
+  return RaftMessage(
+    kind: rmkAppendEntries,
+    term: node.currentTerm,
+    senderId: node.id,
+    prevLogIndex: prevIdx,
+    prevLogTerm: prevTerm,
+    entries: entries,
+    leaderCommit: node.commitIndex,
+  )
+
+proc appendLog*(node: RaftNode, command: string, data: seq[byte] = @[]): LogEntry =
+  if node.state != rsLeader:
+    return LogEntry()
+  result = LogEntry(
+    term: node.currentTerm,
+    index: node.lastLogIndex + 1,
+    command: command,
+    data: data,
+  )
+  node.log.add(result)
+
+proc handleVoteReply*(node: RaftNode, reply: RaftMessage) =
+  if reply.term > node.currentTerm:
+    node.becomeFollower(reply.term)
+    return
+
+  if node.state != rsCandidate:
+    return
+
+  if reply.success:
+    node.votesReceived.incl(reply.senderId)
+    if node.votesReceived.len > (node.peers.len + 1) div 2:
+      node.becomeLeader()
+
+proc handleAppendReply*(node: RaftNode, peerId: string, reply: RaftMessage) =
+  if reply.term > node.currentTerm:
+    node.becomeFollower(reply.term)
+    return
+
+  if node.state != rsLeader:
+    return
+
+  if reply.success:
+    node.matchIndex[peerId] = reply.matchIdx
+    node.nextIndex[peerId] = reply.matchIdx + 1
+
+    # Update commit index
+    var matchIndices: seq[uint64] = @[node.lastLogIndex]
+    for p, idx in node.matchIndex:
+      matchIndices.add(idx)
+    matchIndices.sort()
+
+    let medianIdx = matchIndices[matchIndices.len div 2]
+    if medianIdx > node.commitIndex:
+      if medianIdx <= node.lastLogIndex and
+         node.log[medianIdx - 1].term == node.currentTerm:
+        node.commitIndex = medianIdx
+  else:
+    if node.nextIndex[peerId] > 1:
+      dec node.nextIndex[peerId]
+
+proc state*(node: RaftNode): RaftState = node.state
+proc isLeader*(node: RaftNode): bool = node.state == rsLeader
+proc leaderId*(node: RaftNode): string = node.leaderId
+proc logLen*(node: RaftNode): int = node.log.len

src/barabadb/core/sharding.nim

@@ -0,0 +1,133 @@
+## Sharding — hash-based and range-based data distribution
+import std/tables
+import std/hashes
+import std/algorithm
+import std/sets
+
+type
+  ShardStrategy* = enum
+    ssHash = "hash"
+    ssRange = "range"
+    ssConsistent = "consistent"
+
+  Shard* = object
+    id*: int
+    name*: string
+    minKey*: string
+    maxKey*: string
+    nodeIds*: seq[string]  # replica node ids
+    isActive*: bool
+    entryCount*: int
+
+  ShardRouter* = ref object
+    strategy*: ShardStrategy
+    shards*: seq[Shard]
+    vnodeRing*: seq[(uint64, int)]  # (hash, shard_id) sorted
+    replicas*: int
+
+  ShardConfig* = object
+    numShards*: int
+    replicas*: int
+    strategy*: ShardStrategy
+
+proc defaultShardConfig*(): ShardConfig =
+  ShardConfig(numShards: 4, replicas: 1, strategy: ssHash)
+
+proc newShardRouter*(config: ShardConfig = defaultShardConfig()): ShardRouter =
+  result = ShardRouter(
+    strategy: config.strategy,
+    shards: @[],
+    vnodeRing: @[],
+    replicas: config.replicas,
+  )
+  for i in 0..<config.numShards:
+    result.shards.add(Shard(
+      id: i,
+      name: "shard_" & $i,
+      minKey: "",
+      maxKey: "",
+      nodeIds: @[],
+      isActive: true,
+      entryCount: 0,
+    ))
+
+proc hashKey*(key: string): uint64 =
+  return uint64(hash(key))
+
+proc getShardHash*(router: ShardRouter, key: string): int =
+  let h = hashKey(key)
+  return int(h mod uint64(router.shards.len))
+
+proc getShardRange*(router: ShardRouter, key: string): int =
+  for i, shard in router.shards:
+    if key >= shard.minKey and key <= shard.maxKey:
+      return i
+  return 0
+
+proc getShardConsistent*(router: ShardRouter, key: string): int =
+  if router.vnodeRing.len == 0:
+    return getShardHash(router, key)
+  let h = hashKey(key)
+  for (ringHash, shardId) in router.vnodeRing:
+    if h <= ringHash:
+      return shardId
+  return router.vnodeRing[0][1]
+
+proc getShard*(router: ShardRouter, key: string): int =
+  case router.strategy
+  of ssHash: router.getShardHash(key)
+  of ssRange: router.getShardRange(key)
+  of ssConsistent: router.getShardConsistent(key)
+
+proc addVirtualNodes*(router: var ShardRouter, numVnodes: int = 100) =
+  for shard in router.shards:
+    for i in 0..<numVnodes:
+      let vnodeKey = shard.name & "_vnode_" & $i
+      router.vnodeRing.add((hashKey(vnodeKey), shard.id))
+  router.vnodeRing.sort(proc(a, b: (uint64, int)): int = cmp(a[0], b[0]))
+
+proc setRangeBounds*(router: var ShardRouter, bounds: seq[(string, string)]) =
+  for i, bound in bounds:
+    if i < router.shards.len:
+      router.shards[i].minKey = bound[0]
+      router.shards[i].maxKey = bound[1]
+
+proc assignNode*(router: var ShardRouter, shardId: int, nodeId: string) =
+  if shardId < router.shards.len:
+    router.shards[shardId].nodeIds.add(nodeId)
+
+proc getShardForNode*(router: ShardRouter, nodeId: string): seq[int] =
+  result = @[]
+  for i, shard in router.shards:
+    if nodeId in shard.nodeIds:
+      result.add(i)
+
+proc replicasOf*(router: ShardRouter, key: string): seq[string] =
+  let shardId = router.getShard(key)
+  if shardId < router.shards.len:
+    return router.shards[shardId].nodeIds
+  return @[]
+
+proc allShards*(router: ShardRouter): seq[Shard] =
+  return router.shards
+
+proc activeShardCount*(router: ShardRouter): int =
+  result = 0
+  for s in router.shards:
+    if s.isActive:
+      inc result
+
+proc rebalance*(router: var ShardRouter, nodes: seq[string]) =
+  if nodes.len == 0:
+    return
+  # Clear existing assignments
+  for i in 0..<router.shards.len:
+    router.shards[i].nodeIds = @[]
+
+  # Round-robin assign replicas
+  for i in 0..<router.shards.len:
+    for r in 0..<router.replicas:
+      let nodeIdx = (i + r) mod nodes.len
+      router.shards[i].nodeIds.add(nodes[nodeIdx])
+
+proc shardCount*(router: ShardRouter): int = router.shards.len

src/barabadb/query/ast.nim

@@ -120,7 +120,7 @@ type
     case kind*: NodeKind
     of nkSelect:
       selDistinct*: bool
-      selWith*: seq[Node]
+      selWith*: seq[(string, Node)]
       selResult*: seq[Node]
       selFrom*: Node
       selJoins*: seq[Node]

src/barabadb/query/lexer.nim

@@ -107,6 +107,7 @@ type
     tkAssign
     tkArrow
     tkDoubleColon
+    tkColon
     tkDot
     tkComma
     tkSemicolon
@@ -351,7 +352,7 @@ proc nextToken*(l: var Lexer): Token =
       discard l.advance()
       return Token(kind: tkEq, value: "==", line: startLine, col: startCol)
     discard l.advance()
-    return Token(kind: tkAssign, value: ":=", line: startLine, col: startCol)
+    return Token(kind: tkEq, value: "=", line: startLine, col: startCol)
   of ':':
     if l.pos + 1 < l.input.len and l.input[l.pos + 1] == '=':
       discard l.advance()
@@ -362,7 +363,7 @@ proc nextToken*(l: var Lexer): Token =
       discard l.advance()
       return Token(kind: tkDoubleColon, value: "::", line: startLine, col: startCol)
     discard l.advance()
-    return Token(kind: tkInvalid, value: ":", line: startLine, col: startCol)
+    return Token(kind: tkColon, value: ":", line: startLine, col: startCol)
   of '!':
     if l.pos + 1 < l.input.len and l.input[l.pos + 1] == '=':
       discard l.advance()

src/barabadb/query/parser.nim

@@ -2,6 +2,7 @@
 import std/strutils
 import lexer
 import ast
+import ../core/types
 
 type
   Parser* = object
@@ -56,9 +57,32 @@ proc parsePrimary(p: var Parser): Node =
     Node(kind: nkNullLit, line: tok.line, col: tok.col)
   of tkIdent:
     discard p.advance()
-    Node(kind: nkIdent, identName: tok.value, line: tok.line, col: tok.col)
+    # Check for function call: ident(...)
+    if p.peek().kind == tkLParen:
+      discard p.advance()  # consume (
+      var args: seq[Node] = @[]
+      if p.peek().kind != tkRParen:
+        args.add(p.parseExpr())
+        while p.match(tkComma):
+          args.add(p.parseExpr())
+      discard p.expect(tkRParen)
+      return Node(kind: nkFuncCall, funcName: tok.value, funcArgs: args,
+                  line: tok.line, col: tok.col)
+    # Check for dotted path: ident.ident.ident
+    var parts = @[tok.value]
+    while p.peek().kind == tkDot:
+      discard p.advance()  # consume .
+      parts.add(p.expect(tkIdent).value)
+    if parts.len == 1:
+      return Node(kind: nkIdent, identName: tok.value, line: tok.line, col: tok.col)
+    return Node(kind: nkPath, pathParts: parts, line: tok.line, col: tok.col)
   of tkLParen:
     discard p.advance()
+    # Check for subquery
+    if p.peek().kind == tkSelect:
+      let sub = p.parseSelect()
+      discard p.expect(tkRParen)
+      return Node(kind: nkSubquery, subQuery: sub, line: tok.line, col: tok.col)
     let expr = p.parseExpr()
     discard p.expect(tkRParen)
     expr
@@ -87,14 +111,40 @@ proc parsePrimary(p: var Parser): Node =
     discard p.advance()
     let operand = p.parsePrimary()
     Node(kind: nkUnaryOp, unOp: ukNeg, unOperand: operand, line: tok.line, col: tok.col)
-  of tkCount:
+  of tkCount, tkSum, tkAvg, tkMin, tkMax:
+    let funcName = tok.value
     discard p.advance()
     discard p.expect(tkLParen)
     var args: seq[Node] = @[]
+    # Handle DISTINCT inside aggregate
+    var hasDistinct = false
+    if p.peek().kind == tkDistinct:
+      discard p.advance()
+      hasDistinct = true
     if p.peek().kind != tkRParen:
       args.add(p.parseExpr())
     discard p.expect(tkRParen)
-    Node(kind: nkFuncCall, funcName: "count", funcArgs: args, line: tok.line, col: tok.col)
+    var node = Node(kind: nkFuncCall, funcName: funcName.toLower(), funcArgs: args,
+                    line: tok.line, col: tok.col)
+    return node
+  of tkCase:
+    discard p.advance()
+    var caseExpr: Node = nil
+    # CASE expr WHEN ... THEN ... ELSE ... END
+    if p.peek().kind != tkWhen:
+      caseExpr = p.parseExpr()
+    var whens: seq[(Node, Node)] = @[]
+    while p.match(tkWhen):
+      let cond = p.parseExpr()
+      discard p.expect(tkThen)
+      let val = p.parseExpr()
+      whens.add((cond, val))
+    var elseExpr: Node = nil
+    if p.match(tkElse):
+      elseExpr = p.parseExpr()
+    discard p.expect(tkEnd)
+    Node(kind: nkCase, caseExpr: caseExpr, caseWhens: whens, caseElse: elseExpr,
+         line: tok.line, col: tok.col)
   else:
     discard p.advance()
     Node(kind: nkNullLit, line: tok.line, col: tok.col)
@@ -128,6 +178,37 @@ proc parseAddSub(p: var Parser): Node =
 
 proc parseComparison(p: var Parser): Node =
   result = p.parseAddSub()
+  # Handle BETWEEN ... AND ...
+  if p.peek().kind == tkBetween:
+    let tok = p.advance()
+    let low = p.parseAddSub()
+    discard p.expect(tkAnd)
+    let high = p.parseAddSub()
+    return Node(kind: nkBetweenExpr, betweenExpr: result,
+                betweenLow: low, betweenHigh: high, line: tok.line, col: tok.col)
+  # Handle IN (subquery | list)
+  if p.peek().kind == tkIn:
+    let tok = p.advance()
+    let right = p.parseAddSub()
+    return Node(kind: nkInExpr, inLeft: result, inRight: right,
+                line: tok.line, col: tok.col)
+  # Handle LIKE / ILIKE
+  if p.peek().kind in {tkLike, tkILike}:
+    let isILike = p.peek().kind == tkILike
+    let tok = p.advance()
+    let pattern = p.parseAddSub()
+    return Node(kind: nkLikeExpr, likeExpr: result, likePattern: pattern,
+                likeCaseInsensitive: isILike, line: tok.line, col: tok.col)
+  # Handle IS NULL / IS NOT NULL
+  if p.peek().kind == tkIs:
+    let tok = p.advance()
+    var negated = false
+    if p.peek().kind == tkNot:
+      discard p.advance()
+      negated = true
+    discard p.advance()  # consume NULL token (assumed)
+    return Node(kind: nkIsExpr, isExpr: result, isNegated: negated,
+                line: tok.line, col: tok.col)
   while p.peek().kind in {tkEq, tkNotEq, tkLt, tkLtEq, tkGt, tkGtEq}:
     let op = case p.peek().kind
       of tkEq: bkEq
@@ -169,20 +250,83 @@ proc parseOr(p: var Parser): Node =
 proc parseExpr(p: var Parser): Node =
   return p.parseOr()
 
+proc parseJoinType(p: var Parser): JoinKind =
+  if p.match(tkInner):
+    return jkInner
+  elif p.match(tkLeft):
+    if p.match(tkOuter): discard
+    return jkLeft
+  elif p.match(tkRight):
+    if p.match(tkOuter): discard
+    return jkRight
+  elif p.match(tkFull):
+    if p.match(tkOuter): discard
+    return jkFull
+  elif p.match(tkCross):
+    return jkCross
+  return jkInner
+
+proc parseWith(p: var Parser): Node =
+  # WITH name AS (select), name2 AS (select2) SELECT ...
+  let tok = p.expect(tkWith)
+  result = Node(kind: nkWith, line: tok.line, col: tok.col)
+  result.withBindings = @[]
+
+  # Parse first CTE
+  let cteName = p.expect(tkIdent).value
+  discard p.expect(tkAs)
+  discard p.expect(tkLParen)
+  let cteQuery = p.parseSelect()
+  discard p.expect(tkRParen)
+  result.withBindings.add((cteName, cteQuery))
+
+  # Parse additional CTEs
+  while p.match(tkComma):
+    let name = p.expect(tkIdent).value
+    discard p.expect(tkAs)
+    discard p.expect(tkLParen)
+    let query = p.parseSelect()
+    discard p.expect(tkRParen)
+    result.withBindings.add((name, query))
+
 proc parseSelect(p: var Parser): Node =
+  # Handle WITH (CTE)
+  var withClause: Node = nil
+  if p.peek().kind == tkWith:
+    withClause = p.parseWith()
+
   let tok = p.expect(tkSelect)
   result = Node(kind: nkSelect, line: tok.line, col: tok.col)
 
+  if withClause != nil:
+    result.selWith = withClause.withBindings
+
   if p.peek().kind == tkDistinct:
     discard p.advance()
     result.selDistinct = true
 
+  # Parse SELECT list
   result.selResult = @[]
   result.selResult.add(p.parseExpr())
   while p.match(tkComma):
     result.selResult.add(p.parseExpr())
 
+  # Parse FROM
+  result.selJoins = @[]
   if p.match(tkFrom):
+    # Handle subquery: (SELECT ...) AS alias
+    if p.peek().kind == tkLParen:
+      discard p.advance()  # consume (
+      let subquery = p.parseSelect()
+      discard p.expect(tkRParen)
+      var alias = ""
+      if p.match(tkAs):
+        alias = p.expect(tkIdent).value
+      elif p.peek().kind == tkIdent:
+        alias = p.advance().value
+      result.selFrom = Node(kind: nkFrom, fromTable: "(subquery)",
+                            fromAlias: alias, line: tok.line, col: tok.col)
+    else:
       let tableTok = p.expect(tkIdent)
       var alias = ""
       if p.match(tkAs):
@@ -192,12 +336,70 @@ proc parseSelect(p: var Parser): Node =
       result.selFrom = Node(kind: nkFrom, fromTable: tableTok.value,
                             fromAlias: alias, line: tableTok.line, col: tableTok.col)
 
+    # Parse JOINs
+    while p.peek().kind == tkJoin or
+          (p.peek().kind in {tkInner, tkLeft, tkRight, tkFull, tkCross} and
+           p.pos + 1 < p.tokens.len and p.tokens[p.pos + 1].kind == tkJoin):
+      let jk = p.parseJoinType()
+      discard p.expect(tkJoin)
+      let joinTable = p.expect(tkIdent)
+      var joinAlias = ""
+      if p.match(tkAs):
+        joinAlias = p.expect(tkIdent).value
+      elif p.peek().kind == tkIdent:
+        joinAlias = p.advance().value
+      var joinCond: Node = nil
+      if p.match(tkOn):
+        joinCond = p.parseExpr()
+      let joinTarget = Node(kind: nkFrom, fromTable: joinTable.value,
+                            fromAlias: joinAlias, line: joinTable.line, col: joinTable.col)
+      result.selJoins.add(Node(kind: nkJoin, joinKind: jk, joinTarget: joinTarget,
+                               joinOn: joinCond, joinAlias: joinAlias,
+                               line: joinTable.line, col: joinTable.col))
+
+  # Parse WHERE
   if p.match(tkWhere):
     result.selWhere = Node(kind: nkWhere, whereExpr: p.parseExpr())
 
+  # Parse GROUP BY
+  if p.match(tkGroup):
+    discard p.expect(tkBy)
+    result.selGroupBy = @[]
+    result.selGroupBy.add(p.parseExpr())
+    while p.match(tkComma):
+      result.selGroupBy.add(p.parseExpr())
+
+  # Parse HAVING
+  if p.match(tkHaving):
+    result.selHaving = Node(kind: nkHaving, havingExpr: p.parseExpr())
+
+  # Parse ORDER BY
+  if p.match(tkOrder):
+    discard p.expect(tkBy)
+    result.selOrderBy = @[]
+    var firstExpr = p.parseExpr()
+    var firstDir = sdAsc
+    if p.match(tkDesc):
+      firstDir = sdDesc
+    elif p.match(tkAsc):
+      firstDir = sdAsc
+    result.selOrderBy.add(Node(kind: nkOrderBy, orderByExpr: firstExpr,
+                               orderByDir: firstDir))
+    while p.match(tkComma):
+      let expr = p.parseExpr()
+      var dir = sdAsc
+      if p.match(tkDesc):
+        dir = sdDesc
+      elif p.match(tkAsc):
+        dir = sdAsc
+      result.selOrderBy.add(Node(kind: nkOrderBy, orderByExpr: expr,
+                                 orderByDir: dir))
+
+  # Parse LIMIT
   if p.match(tkLimit):
     result.selLimit = Node(kind: nkLimit, limitExpr: p.parseExpr())
 
+  # Parse OFFSET
   if p.match(tkOffset):
     result.selOffset = Node(kind: nkOffset, offsetExpr: p.parseExpr())
 
@@ -210,21 +412,81 @@ proc parseUpdate(p: var Parser): Node =
   let tok = p.expect(tkUpdate)
   let target = p.expect(tkIdent).value
   result = Node(kind: nkUpdate, updTarget: target, line: tok.line, col: tok.col)
+  if p.match(tkSet):
+    result.updSet = @[]
+    let field = p.expect(tkIdent).value
+    discard p.match(tkEq)  # = or :=
+    let val = p.parseExpr()
+    result.updSet.add(Node(kind: nkBinOp, binOp: bkAssign,
+      binLeft: Node(kind: nkIdent, identName: field),
+      binRight: val))
+    while p.match(tkComma):
+      let f = p.expect(tkIdent).value
+      discard p.match(tkEq)
+      let v = p.parseExpr()
+      result.updSet.add(Node(kind: nkBinOp, binOp: bkAssign,
+        binLeft: Node(kind: nkIdent, identName: f),
+        binRight: v))
+  if p.match(tkWhere):
+    result.updWhere = Node(kind: nkWhere, whereExpr: p.parseExpr())
 
 proc parseDelete(p: var Parser): Node =
   let tok = p.expect(tkDelete)
+  discard p.match(tkFrom)  # optional FROM keyword
   let target = p.expect(tkIdent).value
   result = Node(kind: nkDelete, delTarget: target, line: tok.line, col: tok.col)
+  if p.match(tkWhere):
+    result.delWhere = Node(kind: nkWhere, whereExpr: p.parseExpr())
 
 proc parseCreateType(p: var Parser): Node =
   let tok = p.expect(tkCreate)
   discard p.expect(tkType)
   let name = p.expect(tkIdent).value
   result = Node(kind: nkCreateType, ctName: name, line: tok.line, col: tok.col)
+  # Parse bases (EXTENDING)
+  result.ctBases = @[]
+  if p.match(tkIdent):  # "extending" keyword mapped to ident
+    # Check if the ident is "extending"
+    # For now, just accept bases in braces
+    discard
+  # Parse body
+  if p.match(tkLBrace):
+    result.ctProperties = @[]
+    result.ctLinks = @[]
+    while p.peek().kind != tkRbrace and p.peek().kind != tkEof:
+      discard p.match(tkComma)  # optional comma separator
+      # Parse property or link
+      var isRequired = false
+      var isMulti = false
+      if p.peek().kind == tkRequired:
+        discard p.advance()
+        isRequired = true
+      if p.peek().kind == tkMulti:
+        discard p.advance()
+        isMulti = true
+
+      let fieldTok = p.expect(tkIdent)
+      # Check for link or property
+      if p.peek().kind == tkArrow:  # -> means link
+        discard p.advance()  # consume ->
+        let target = p.expect(tkIdent).value
+        result.ctLinks.add(Node(kind: nkLinkDef,
+          ldName: fieldTok.value, ldTarget: target,
+          ldRequired: isRequired,
+          ldCardinality: if isMulti: Many else: One))
+      else:
+        var typeName = ""
+        if p.match(tkColon):
+          typeName = p.expect(tkIdent).value
+        result.ctProperties.add(Node(kind: nkPropertyDef,
+          pdName: fieldTok.value, pdType: typeName,
+          pdRequired: isRequired))
+      discard p.match(tkSemicolon)
+    discard p.expect(tkRbrace)
 
 proc parseStatement*(p: var Parser): Node =
   case p.peek().kind
-  of tkSelect: p.parseSelect()
+  of tkWith, tkSelect: p.parseSelect()
   of tkInsert: p.parseInsert()
   of tkUpdate: p.parseUpdate()
   of tkDelete: p.parseDelete()

tests/test_all.nim

@@ -7,6 +7,8 @@ import barabadb/core/types
 import barabadb/core/mvcc
 import barabadb/core/deadlock
 import barabadb/core/columnar
+import barabadb/core/raft
+import barabadb/core/sharding
 import barabadb/storage/bloom
 import barabadb/storage/wal
 import barabadb/storage/lsm
@@ -782,3 +784,173 @@ suite "Vector Metadata Filtering":
     let results = vengine.searchWithFilter(idx, @[1.0'f32, 0.0'f32, 0.0'f32], 10,
                                            filter = filterA)
     check results.len == 2  # only category A entries
+
+suite "BaraQL Parser — Extended":
+  test "Parse GROUP BY":
+    let ast = parse("SELECT dept, count(*) FROM employees GROUP BY dept")
+    check ast.stmts.len == 1
+    check ast.stmts[0].selGroupBy.len == 1
+
+  test "Parse GROUP BY with HAVING":
+    let ast = parse("SELECT dept, count(*) FROM employees GROUP BY dept HAVING count(*) > 5")
+    check ast.stmts[0].selHaving != nil
+
+  test "Parse ORDER BY with direction":
+    let ast = parse("SELECT name FROM users ORDER BY age DESC")
+    check ast.stmts[0].selOrderBy.len == 1
+    check ast.stmts[0].selOrderBy[0].orderByDir == sdDesc
+
+  test "Parse ORDER BY multiple columns":
+    let ast = parse("SELECT * FROM t ORDER BY a ASC, b DESC")
+    check ast.stmts[0].selOrderBy.len == 2
+
+  test "Parse INNER JOIN":
+    let ast = parse("SELECT u.name, o.total FROM users u INNER JOIN orders o ON u.id = o.user_id")
+    check ast.stmts[0].selJoins.len == 1
+    check ast.stmts[0].selJoins[0].joinKind == jkInner
+
+  test "Parse LEFT JOIN":
+    let ast = parse("SELECT u.name FROM users u LEFT JOIN orders o ON u.id = o.user_id")
+    check ast.stmts[0].selJoins.len == 1
+    check ast.stmts[0].selJoins[0].joinKind == jkLeft
+
+  test "Parse multiple JOINs":
+    let ast = parse("SELECT * FROM a JOIN b ON a.id = b.aid JOIN c ON b.id = c.bid")
+    check ast.stmts[0].selJoins.len == 2
+
+  test "Parse CTE (WITH)":
+    let ast = parse("WITH active AS (SELECT * FROM users WHERE active = true) SELECT * FROM active")
+    check ast.stmts[0].selWith.len == 1
+    check ast.stmts[0].selWith[0][0] == "active"
+
+  test "Parse multiple CTEs":
+    let ast = parse("WITH a AS (SELECT id FROM t1), b AS (SELECT id FROM t2) SELECT * FROM a")
+    check ast.stmts[0].selWith.len == 2
+
+  test "Parse aggregate functions in SELECT":
+    let ast = parse("SELECT count(*), sum(amount), avg(price), min(age), max(score) FROM orders")
+    check ast.stmts[0].selResult.len == 5
+
+  test "Parse CASE expression":
+    let ast = parse("SELECT CASE WHEN age > 18 THEN 'adult' ELSE 'minor' END FROM users")
+    check ast.stmts[0].selResult.len == 1
+
+  test "Parse BETWEEN":
+    let ast = parse("SELECT * FROM products WHERE price BETWEEN 10 AND 100")
+    check ast.stmts[0].selWhere != nil
+
+  test "Parse subquery in FROM":
+    let ast = parse("SELECT * FROM (SELECT id FROM users) AS sub")
+    check ast.stmts[0].selFrom != nil
+
+  test "Parse UPDATE SET WHERE":
+    let ast = parse("UPDATE users SET name = 'Alice' WHERE id = 1")
+    check ast.stmts[0].updSet.len == 1
+
+  test "Parse DELETE WHERE":
+    let ast = parse("DELETE FROM users WHERE id = 1")
+    check ast.stmts[0].delWhere != nil
+
+  test "Parse CREATE TYPE with properties":
+    let ast = parse("CREATE TYPE Person { name: str, age: int32 }")
+    check ast.stmts[0].ctName == "Person"
+    check ast.stmts[0].ctProperties.len == 2
+
+suite "Raft Consensus":
+  test "Create cluster with nodes":
+    var cluster = newRaftCluster()
+    cluster.addNode("n1")
+    cluster.addNode("n2")
+    cluster.addNode("n3")
+    check cluster.nodes.len == 3
+    check cluster.nodes["n1"].peers.len == 2
+
+  test "Initial state is follower":
+    var cluster = newRaftCluster()
+    cluster.addNode("n1")
+    check cluster.nodes["n1"].state == rsFollower
+
+  test "Election — single node becomes leader":
+    var cluster = newRaftCluster()
+    cluster.addNode("n1")
+    let node = cluster.nodes["n1"]
+    node.becomeCandidate()
+    node.becomeLeader()
+    check node.isLeader
+    check node.leaderId == "n1"
+
+  test "Log replication":
+    var cluster = newRaftCluster()
+    cluster.addNode("n1")
+    let node = cluster.nodes["n1"]
+    node.becomeCandidate()
+    node.becomeLeader()
+    let entry = node.appendLog("SET key1 value1")
+    check entry.index == 1
+    check node.logLen == 1
+    let entry2 = node.appendLog("SET key2 value2")
+    check entry2.index == 2
+    check node.logLen == 2
+
+  test "RequestVote handling":
+    var cluster = newRaftCluster()
+    cluster.addNode("n1")
+    cluster.addNode("n2")
+    let n1 = cluster.nodes["n1"]
+    let n2 = cluster.nodes["n2"]
+    let req = RaftMessage(kind: rmkRequestVote, term: 1, senderId: "n2",
+                          lastLogIndex: 0, lastLogTerm: 0)
+    let reply = n1.handleRequestVote(req)
+    check reply.success
+    check n1.votedFor == "n2"
+
+  test "AppendEntries handling":
+    var cluster = newRaftCluster()
+    cluster.addNode("n1")
+    cluster.addNode("n2")
+    let n2 = cluster.nodes["n2"]
+    let msg = RaftMessage(kind: rmkAppendEntries, term: 1, senderId: "n1",
+                          prevLogIndex: 0, prevLogTerm: 0,
+                          entries: @[LogEntry(term: 1, index: 1, command: "SET x 1")],
+                          leaderCommit: 0)
+    let reply = n2.handleAppendEntries(msg)
+    check reply.success
+    check n2.logLen == 1
+
+suite "Sharding":
+  test "Hash-based sharding":
+    var router = newShardRouter(ShardConfig(numShards: 4, strategy: ssHash))
+    check router.shardCount == 4
+    let s1 = router.getShard("user_1")
+    let s2 = router.getShard("user_2")
+    check s1 >= 0 and s1 < 4
+    check s2 >= 0 and s2 < 4
+
+  test "Consistent hashing":
+    var router = newShardRouter(ShardConfig(numShards: 4, strategy: ssConsistent))
+    router.addVirtualNodes(50)
+    let s = router.getShard("some_key")
+    check s >= 0 and s < 4
+
+  test "Range-based sharding":
+    var router = newShardRouter(ShardConfig(numShards: 3, strategy: ssRange))
+    router.setRangeBounds(@[("a", "f"), ("g", "n"), ("o", "z")])
+    check router.getShardRange("apple") == 0
+    check router.getShardRange("hello") == 1
+    check router.getShardRange("top") == 2
+
+  test "Rebalance assigns nodes":
+    var router = newShardRouter(ShardConfig(numShards: 3, replicas: 2, strategy: ssHash))
+    router.rebalance(@["node1", "node2", "node3"])
+    for shard in router.shards:
+      check shard.nodeIds.len == 2  # 2 replicas
+
+  test "Replicas of key":
+    var router = newShardRouter(ShardConfig(numShards: 2, replicas: 1, strategy: ssHash))
+    router.rebalance(@["n1", "n2"])
+    let replicas = router.replicasOf("test_key")
+    check replicas.len == 1
+
+  test "Active shard count":
+    var router = newShardRouter()
+    check router.activeShardCount == 4