path: root/libpanto/src/anthropic_messages_json.zig

//! Anthropic Messages API JSON serialization and event parsing.
//!
//! Two responsibilities:
//!   1. Serialize a `Conversation` into a `/v1/messages` request body.
//!      Anthropic differs from OpenAI in several ways — see `serializeRequest`.
//!   2. Parse one streaming SSE event payload (the JSON object after `data: `)
//!      into a strongly-typed `StreamEvent` that the provider can consume.
//!
//! Wire format reference:
//!   https://platform.claude.com/docs/en/build-with-claude/streaming
//!   https://platform.claude.com/docs/en/build-with-claude/extended-thinking

const std = @import("std");
const Allocator = std.mem.Allocator;
const Writer = std.Io.Writer;
const conversation = @import("conversation.zig");
const config_mod = @import("config.zig");

// -----------------------------------------------------------------------------
// Request serialization
// -----------------------------------------------------------------------------

/// Serialize a Conversation into a `/v1/messages` request body.
///
/// Differences from OpenAI Chat Completions:
///   - System messages are extracted and concatenated into a top-level
///     `system` string. They do not appear in the `messages` array.
///   - `content` is always an array of typed blocks, never a bare string.
///   - `max_tokens` is required.
///
/// Caller owns the returned slice.
pub fn serializeRequest(
    allocator: Allocator,
    cfg: *const config_mod.AnthropicMessagesConfig,
    conv: *const conversation.Conversation,
) ![]u8 {
    var aw: Writer.Allocating = .init(allocator);
    errdefer aw.deinit();

    var s: std.json.Stringify = .{ .writer = &aw.writer };

    try s.beginObject();

    try s.objectField("model");
    try s.write(cfg.model);

    try s.objectField("max_tokens");
    try s.write(cfg.max_tokens);

    try s.objectField("stream");
    try s.write(true);

    // Build and emit the concatenated system prompt, if any.
    var system_buf: std.ArrayList(u8) = .empty;
    defer system_buf.deinit(allocator);
    try collectSystemPrompt(conv, &system_buf, allocator);
    if (system_buf.items.len > 0) {
        try s.objectField("system");
        try s.write(system_buf.items);
    }

    // Emit messages (everything that isn't .system).
    try s.objectField("messages");
    try s.beginArray();
    for (conv.messages.items) |msg| {
        if (msg.role == .system) continue;
        try writeMessage(&s, msg, allocator);
    }
    try s.endArray();

    try s.endObject();

    return try aw.toOwnedSlice();
}

fn collectSystemPrompt(
    conv: *const conversation.Conversation,
    out: *std.ArrayList(u8),
    allocator: Allocator,
) !void {
    var first = true;
    for (conv.messages.items) |msg| {
        if (msg.role != .system) continue;
        for (msg.content.items) |block| {
            if (block != .Text) continue;
            if (!first) try out.append(allocator, '\n');
            try out.appendSlice(allocator, block.Text.items);
            first = false;
        }
    }
}

fn writeMessage(
    s: *std.json.Stringify,
    msg: conversation.Message,
    allocator: Allocator,
) !void {
    _ = allocator;
    try s.beginObject();

    try s.objectField("role");
    try s.write(@tagName(msg.role));

    try s.objectField("content");
    try s.beginArray();
    for (msg.content.items) |block| {
        try writeBlock(s, block);
    }
    try s.endArray();

    try s.endObject();
}

fn writeBlock(s: *std.json.Stringify, block: conversation.ContentBlock) !void {
    switch (block) {
        .Text => |tb| {
            try s.beginObject();
            try s.objectField("type");
            try s.write("text");
            try s.objectField("text");
            try s.write(tb.items);
            try s.endObject();
        },
        .Thinking => |tb| {
            // Anthropic requires the signature field to round-trip a thinking
            // block. If we don't have one (e.g. block was synthesized from a
            // non-Anthropic provider), skip the block: Anthropic will reject
            // unsigned thinking on incoming messages.
            const sig = tb.signature orelse return;
            try s.beginObject();
            try s.objectField("type");
            try s.write("thinking");
            try s.objectField("thinking");
            try s.write(tb.text.items);
            try s.objectField("signature");
            try s.write(sig);
            try s.endObject();
        },
        // ToolUse / ToolResult: phase 3+.
        .ToolUse, .ToolResult => {},
    }
}

// -----------------------------------------------------------------------------
// Streaming event parsing
// -----------------------------------------------------------------------------

/// Parsed shape of one SSE event payload. The `tag` field selects which of
/// the payload variants is populated.
///
/// String slices are borrowed from the underlying `std.json.Parsed`. The
/// caller must keep `parsed` alive (or copy the slices) until the event is
/// fully consumed.
pub const StreamEventTag = enum {
    message_start,
    content_block_start,
    content_block_delta,
    content_block_stop,
    message_delta,
    message_stop,
    ping,
    @"error",
    unknown,
};

pub const ContentBlockKind = enum { text, thinking, tool_use, unknown };

pub const StreamEvent = union(StreamEventTag) {
    message_start: struct {
        role: ?[]const u8 = null,
    },
    content_block_start: struct {
        index: usize,
        kind: ContentBlockKind,
        // For tool_use blocks (phase 3+):
        tool_id: ?[]const u8 = null,
        tool_name: ?[]const u8 = null,
    },
    content_block_delta: struct {
        index: usize,
        text_delta: ?[]const u8 = null,
        thinking_delta: ?[]const u8 = null,
        signature_delta: ?[]const u8 = null,
        input_json_delta: ?[]const u8 = null,
    },
    content_block_stop: struct {
        index: usize,
    },
    message_delta: struct {
        stop_reason: ?[]const u8 = null,
    },
    message_stop: void,
    ping: void,
    @"error": struct {
        kind: ?[]const u8 = null,
        message: ?[]const u8 = null,
    },
    unknown: void,
};

pub const ParsedStreamEvent = struct {
    parsed: std.json.Parsed(std.json.Value),
    event: StreamEvent,

    pub fn deinit(self: *ParsedStreamEvent) void {
        self.parsed.deinit();
    }
};

pub fn parseStreamEvent(allocator: Allocator, payload: []const u8) !ParsedStreamEvent {
    var parsed = try std.json.parseFromSlice(std.json.Value, allocator, payload, .{});
    errdefer parsed.deinit();

    const root = parsed.value;
    if (root != .object) return .{ .parsed = parsed, .event = .unknown };

    const type_v = root.object.get("type") orelse return .{ .parsed = parsed, .event = .unknown };
    if (type_v != .string) return .{ .parsed = parsed, .event = .unknown };

    const ty = type_v.string;
    if (std.mem.eql(u8, ty, "message_start")) {
        var role: ?[]const u8 = null;
        if (root.object.get("message")) |m| {
            if (m == .object) {
                if (m.object.get("role")) |r| {
                    if (r == .string) role = r.string;
                }
            }
        }
        return .{ .parsed = parsed, .event = .{ .message_start = .{ .role = role } } };
    }

    if (std.mem.eql(u8, ty, "content_block_start")) {
        const idx = readIndex(root) orelse 0;
        var kind: ContentBlockKind = .unknown;
        var tool_id: ?[]const u8 = null;
        var tool_name: ?[]const u8 = null;
        if (root.object.get("content_block")) |cb| {
            if (cb == .object) {
                if (cb.object.get("type")) |t| {
                    if (t == .string) {
                        if (std.mem.eql(u8, t.string, "text")) {
                            kind = .text;
                        } else if (std.mem.eql(u8, t.string, "thinking")) {
                            kind = .thinking;
                        } else if (std.mem.eql(u8, t.string, "tool_use")) {
                            kind = .tool_use;
                        }
                    }
                }
                if (cb.object.get("id")) |i| {
                    if (i == .string) tool_id = i.string;
                }
                if (cb.object.get("name")) |n| {
                    if (n == .string) tool_name = n.string;
                }
            }
        }
        return .{ .parsed = parsed, .event = .{ .content_block_start = .{
            .index = idx,
            .kind = kind,
            .tool_id = tool_id,
            .tool_name = tool_name,
        } } };
    }

    if (std.mem.eql(u8, ty, "content_block_delta")) {
        const idx = readIndex(root) orelse 0;
        var text_delta: ?[]const u8 = null;
        var thinking_delta: ?[]const u8 = null;
        var signature_delta: ?[]const u8 = null;
        var input_json_delta: ?[]const u8 = null;
        if (root.object.get("delta")) |d| {
            if (d == .object) {
                const dt = blk: {
                    if (d.object.get("type")) |t| {
                        if (t == .string) break :blk t.string;
                    }
                    break :blk "";
                };
                if (std.mem.eql(u8, dt, "text_delta")) {
                    if (d.object.get("text")) |v| if (v == .string) {
                        text_delta = v.string;
                    };
                } else if (std.mem.eql(u8, dt, "thinking_delta")) {
                    if (d.object.get("thinking")) |v| if (v == .string) {
                        thinking_delta = v.string;
                    };
                } else if (std.mem.eql(u8, dt, "signature_delta")) {
                    if (d.object.get("signature")) |v| if (v == .string) {
                        signature_delta = v.string;
                    };
                } else if (std.mem.eql(u8, dt, "input_json_delta")) {
                    if (d.object.get("partial_json")) |v| if (v == .string) {
                        input_json_delta = v.string;
                    };
                }
            }
        }
        return .{ .parsed = parsed, .event = .{ .content_block_delta = .{
            .index = idx,
            .text_delta = text_delta,
            .thinking_delta = thinking_delta,
            .signature_delta = signature_delta,
            .input_json_delta = input_json_delta,
        } } };
    }

    if (std.mem.eql(u8, ty, "content_block_stop")) {
        const idx = readIndex(root) orelse 0;
        return .{ .parsed = parsed, .event = .{ .content_block_stop = .{ .index = idx } } };
    }

    if (std.mem.eql(u8, ty, "message_delta")) {
        var stop_reason: ?[]const u8 = null;
        if (root.object.get("delta")) |d| {
            if (d == .object) {
                if (d.object.get("stop_reason")) |sr| {
                    if (sr == .string) stop_reason = sr.string;
                }
            }
        }
        return .{ .parsed = parsed, .event = .{ .message_delta = .{ .stop_reason = stop_reason } } };
    }

    if (std.mem.eql(u8, ty, "message_stop")) {
        return .{ .parsed = parsed, .event = .message_stop };
    }

    if (std.mem.eql(u8, ty, "ping")) {
        return .{ .parsed = parsed, .event = .ping };
    }

    if (std.mem.eql(u8, ty, "error")) {
        var kind: ?[]const u8 = null;
        var message: ?[]const u8 = null;
        if (root.object.get("error")) |e| {
            if (e == .object) {
                if (e.object.get("type")) |t| if (t == .string) {
                    kind = t.string;
                };
                if (e.object.get("message")) |m| if (m == .string) {
                    message = m.string;
                };
            }
        }
        return .{ .parsed = parsed, .event = .{ .@"error" = .{ .kind = kind, .message = message } } };
    }

    return .{ .parsed = parsed, .event = .unknown };
}

fn readIndex(root: std.json.Value) ?usize {
    const v = root.object.get("index") orelse return null;
    if (v != .integer) return null;
    if (v.integer < 0) return null;
    return @intCast(v.integer);
}

// -----------------------------------------------------------------------------
// Tests
// -----------------------------------------------------------------------------

const testing = std.testing;

fn testConfig(model: []const u8) config_mod.AnthropicMessagesConfig {
    return .{
        .api_key = "k",
        .base_url = "u",
        .model = model,
        .max_tokens = 1024,
    };
}

test "serializeRequest - system extracted into top-level field" {
    const allocator = testing.allocator;

    var conv = conversation.Conversation.init(allocator);
    defer conv.deinit();

    try conv.addSystemMessage("You are helpful.");
    try conv.addUserMessage("Hello!");

    const cfg = testConfig("claude-sonnet-4-20250514");
    const body = try serializeRequest(allocator, &cfg, &conv);
    defer allocator.free(body);

    var parsed = try std.json.parseFromSlice(std.json.Value, allocator, body, .{});
    defer parsed.deinit();

    const root = parsed.value.object;
    try testing.expectEqualStrings("claude-sonnet-4-20250514", root.get("model").?.string);
    try testing.expect(root.get("stream").?.bool);
    try testing.expectEqual(@as(i64, 1024), root.get("max_tokens").?.integer);
    try testing.expectEqualStrings("You are helpful.", root.get("system").?.string);

    // Only the user message appears in `messages`.
    const msgs = root.get("messages").?.array.items;
    try testing.expectEqual(@as(usize, 1), msgs.len);
    try testing.expectEqualStrings("user", msgs[0].object.get("role").?.string);

    // Content is an array of typed blocks, never a bare string.
    const content = msgs[0].object.get("content").?.array.items;
    try testing.expectEqual(@as(usize, 1), content.len);
    try testing.expectEqualStrings("text", content[0].object.get("type").?.string);
    try testing.expectEqualStrings("Hello!", content[0].object.get("text").?.string);
}

test "serializeRequest - multiple system messages concatenated with newlines" {
    const allocator = testing.allocator;

    var conv = conversation.Conversation.init(allocator);
    defer conv.deinit();

    try conv.addSystemMessage("Be terse.");
    try conv.addSystemMessage("Be accurate.");
    try conv.addUserMessage("Hi");

    const cfg = testConfig("claude-x");
    const body = try serializeRequest(allocator, &cfg, &conv);
    defer allocator.free(body);

    var parsed = try std.json.parseFromSlice(std.json.Value, allocator, body, .{});
    defer parsed.deinit();
    try testing.expectEqualStrings(
        "Be terse.\nBe accurate.",
        parsed.value.object.get("system").?.string,
    );
}

test "serializeRequest - no system messages omits the system field" {
    const allocator = testing.allocator;

    var conv = conversation.Conversation.init(allocator);
    defer conv.deinit();
    try conv.addUserMessage("Hi");

    const cfg = testConfig("claude-x");
    const body = try serializeRequest(allocator, &cfg, &conv);
    defer allocator.free(body);

    var parsed = try std.json.parseFromSlice(std.json.Value, allocator, body, .{});
    defer parsed.deinit();
    try testing.expect(parsed.value.object.get("system") == null);
}

test "serializeRequest - signed assistant Thinking blocks round-trip" {
    const allocator = testing.allocator;

    var conv = conversation.Conversation.init(allocator);
    defer conv.deinit();

    const sig = try allocator.dupe(u8, "EqQBCgIYAhIM1gbcDa9GJwZA");
    try conv.addAssistantMessage(&.{
        .{ .Thinking = .{
            .text = try conversation.textualBlockFromSlice(allocator, "let me think"),
            .signature = sig,
        } },
        .{ .Text = try conversation.textualBlockFromSlice(allocator, "the answer is 42") },
    });

    const cfg = testConfig("claude-x");
    const body = try serializeRequest(allocator, &cfg, &conv);
    defer allocator.free(body);

    var parsed = try std.json.parseFromSlice(std.json.Value, allocator, body, .{});
    defer parsed.deinit();

    const msg = parsed.value.object.get("messages").?.array.items[0];
    try testing.expectEqualStrings("assistant", msg.object.get("role").?.string);

    const content = msg.object.get("content").?.array.items;
    try testing.expectEqual(@as(usize, 2), content.len);

    try testing.expectEqualStrings("thinking", content[0].object.get("type").?.string);
    try testing.expectEqualStrings("let me think", content[0].object.get("thinking").?.string);
    try testing.expectEqualStrings(
        "EqQBCgIYAhIM1gbcDa9GJwZA",
        content[0].object.get("signature").?.string,
    );

    try testing.expectEqualStrings("text", content[1].object.get("type").?.string);
    try testing.expectEqualStrings("the answer is 42", content[1].object.get("text").?.string);
}

test "serializeRequest - unsigned Thinking blocks are dropped" {
    // Anthropic rejects thinking blocks without a valid signature on inbound
    // messages. If a block lacks one (e.g. from a different provider or an
    // interrupted stream), omit it rather than send a guaranteed-400 request.
    const allocator = testing.allocator;

    var conv = conversation.Conversation.init(allocator);
    defer conv.deinit();

    try conv.addAssistantMessage(&.{
        .{ .Thinking = .{
            .text = try conversation.textualBlockFromSlice(allocator, "unsigned thinking"),
            .signature = null,
        } },
        .{ .Text = try conversation.textualBlockFromSlice(allocator, "answer") },
    });

    const cfg = testConfig("claude-x");
    const body = try serializeRequest(allocator, &cfg, &conv);
    defer allocator.free(body);

    var parsed = try std.json.parseFromSlice(std.json.Value, allocator, body, .{});
    defer parsed.deinit();

    const content = parsed.value.object.get("messages").?.array.items[0]
        .object.get("content").?.array.items;
    try testing.expectEqual(@as(usize, 1), content.len);
    try testing.expectEqualStrings("text", content[0].object.get("type").?.string);
}

test "parseStreamEvent - message_start" {
    const allocator = testing.allocator;
    const payload =
        \\{"type":"message_start","message":{"id":"msg_1","type":"message","role":"assistant","content":[],"model":"claude","stop_reason":null}}
    ;
    var pe = try parseStreamEvent(allocator, payload);
    defer pe.deinit();
    try testing.expectEqual(StreamEventTag.message_start, @as(StreamEventTag, pe.event));
    try testing.expectEqualStrings("assistant", pe.event.message_start.role.?);
}

test "parseStreamEvent - content_block_start text" {
    const allocator = testing.allocator;
    const payload =
        \\{"type":"content_block_start","index":0,"content_block":{"type":"text","text":""}}
    ;
    var pe = try parseStreamEvent(allocator, payload);
    defer pe.deinit();
    try testing.expectEqual(StreamEventTag.content_block_start, @as(StreamEventTag, pe.event));
    try testing.expectEqual(@as(usize, 0), pe.event.content_block_start.index);
    try testing.expectEqual(ContentBlockKind.text, pe.event.content_block_start.kind);
}

test "parseStreamEvent - content_block_start thinking" {
    const allocator = testing.allocator;
    const payload =
        \\{"type":"content_block_start","index":0,"content_block":{"type":"thinking","thinking":"","signature":""}}
    ;
    var pe = try parseStreamEvent(allocator, payload);
    defer pe.deinit();
    try testing.expectEqual(ContentBlockKind.thinking, pe.event.content_block_start.kind);
}

test "parseStreamEvent - text_delta" {
    const allocator = testing.allocator;
    const payload =
        \\{"type":"content_block_delta","index":1,"delta":{"type":"text_delta","text":"Hello"}}
    ;
    var pe = try parseStreamEvent(allocator, payload);
    defer pe.deinit();
    try testing.expectEqual(@as(usize, 1), pe.event.content_block_delta.index);
    try testing.expectEqualStrings("Hello", pe.event.content_block_delta.text_delta.?);
    try testing.expect(pe.event.content_block_delta.thinking_delta == null);
}

test "parseStreamEvent - thinking_delta" {
    const allocator = testing.allocator;
    const payload =
        \\{"type":"content_block_delta","index":0,"delta":{"type":"thinking_delta","thinking":"step 1"}}
    ;
    var pe = try parseStreamEvent(allocator, payload);
    defer pe.deinit();
    try testing.expectEqualStrings("step 1", pe.event.content_block_delta.thinking_delta.?);
}

test "parseStreamEvent - signature_delta" {
    const allocator = testing.allocator;
    const payload =
        \\{"type":"content_block_delta","index":0,"delta":{"type":"signature_delta","signature":"abc123"}}
    ;
    var pe = try parseStreamEvent(allocator, payload);
    defer pe.deinit();
    try testing.expectEqualStrings("abc123", pe.event.content_block_delta.signature_delta.?);
}

test "parseStreamEvent - content_block_stop" {
    const allocator = testing.allocator;
    const payload =
        \\{"type":"content_block_stop","index":2}
    ;
    var pe = try parseStreamEvent(allocator, payload);
    defer pe.deinit();
    try testing.expectEqual(@as(usize, 2), pe.event.content_block_stop.index);
}

test "parseStreamEvent - message_delta stop_reason" {
    const allocator = testing.allocator;
    const payload =
        \\{"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null}}
    ;
    var pe = try parseStreamEvent(allocator, payload);
    defer pe.deinit();
    try testing.expectEqualStrings("end_turn", pe.event.message_delta.stop_reason.?);
}

test "parseStreamEvent - message_stop" {
    const allocator = testing.allocator;
    const payload =
        \\{"type":"message_stop"}
    ;
    var pe = try parseStreamEvent(allocator, payload);
    defer pe.deinit();
    try testing.expectEqual(StreamEventTag.message_stop, @as(StreamEventTag, pe.event));
}

test "parseStreamEvent - ping" {
    const allocator = testing.allocator;
    const payload = "{\"type\":\"ping\"}";
    var pe = try parseStreamEvent(allocator, payload);
    defer pe.deinit();
    try testing.expectEqual(StreamEventTag.ping, @as(StreamEventTag, pe.event));
}

test "parseStreamEvent - error" {
    const allocator = testing.allocator;
    const payload =
        \\{"type":"error","error":{"type":"overloaded_error","message":"too busy"}}
    ;
    var pe = try parseStreamEvent(allocator, payload);
    defer pe.deinit();
    try testing.expectEqualStrings("overloaded_error", pe.event.@"error".kind.?);
    try testing.expectEqualStrings("too busy", pe.event.@"error".message.?);
}

test "parseStreamEvent - unknown type" {
    const allocator = testing.allocator;
    const payload =
        \\{"type":"some_future_event","extra":"data"}
    ;
    var pe = try parseStreamEvent(allocator, payload);
    defer pe.deinit();
    try testing.expectEqual(StreamEventTag.unknown, @as(StreamEventTag, pe.event));
}