James WamplerandClaude Opus 5
0d7b7a6f30
Builds out the vertical slice for planning and writing a novel. Three front ends — the React UI, an embedded Claude agent, and an MCP stdio server — all go through one REST API, so an edit made from Claude Code and one made in the browser are the same edit. Layout: Domain entities and enums, no dependencies Application services, DTOs, the agent tool-use loop and its 15 tools Infrastructure EF Core 10 + SQLite, Anthropic SDK client Api ASP.NET Core 10 minimal APIs, OpenAPI, ProblemDetails Mcp MCP stdio server, 21 tools over the same REST API Web React 19 + Vite + TanStack Query + Tailwind v4 Data model is Project > Characters / OutlineNodes / Chapters > Scenes, plus agent conversations. The outline is a self-nesting tree so acts, sequences and beats can be arranged however the book wants; scenes carry goal/conflict/outcome because that is what the agent drafts prose from. Notes on a few choices: - Conversation history replays to the model as text only. The agent re-reads current state through its tools rather than trusting a record of edits that may since have changed in the UI. - The user's turn is persisted before the tool loop runs, so a question is recorded even when the model call fails. Turn order uses an explicit sequence column; timestamps tie when a turn completes inside one tick. - Tool failures return is_error results rather than throwing, so the model can read the message and correct itself. MCP tools do the same via CallToolResult, which keeps the API's own message instead of a generic SDK error. - The Anthropic client is constructed lazily. It is injected into the agent service, which also serves read-only endpoints, and those should keep working on an install with no key. Sending without one returns 503, not 400. - DateTimeOffset is stored as UTC ticks. SQLite refuses to ORDER BY the default text form, which every "recently updated first" listing depends on. Tests run against real in-memory SQLite rather than the EF in-memory provider so they exercise the cascade deletes and query translation that actually ship. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01S56bfZMGe1hnhpWP4CjjNw
48 lines
1.7 KiB
C#
48 lines
1.7 KiB
C#
namespace NovelSoftware.Domain.Entities;
|
|
|
|
/// <summary>A chat thread between the writer and the embedded agent, scoped to one project.</summary>
|
|
public class AgentConversation
|
|
{
|
|
public Guid Id { get; set; } = Guid.NewGuid();
|
|
public Guid ProjectId { get; set; }
|
|
public Project? Project { get; set; }
|
|
|
|
public string Title { get; set; } = "New conversation";
|
|
|
|
public DateTimeOffset CreatedAt { get; set; } = DateTimeOffset.UtcNow;
|
|
public DateTimeOffset UpdatedAt { get; set; } = DateTimeOffset.UtcNow;
|
|
|
|
public List<AgentMessage> Messages { get; set; } = [];
|
|
}
|
|
|
|
/// <summary>
|
|
/// One turn in an agent conversation. Assistant turns may carry a record of the tools
|
|
/// the agent called, so the UI can show what it changed and the next request can replay
|
|
/// the turn back to the model.
|
|
/// </summary>
|
|
public class AgentMessage
|
|
{
|
|
public Guid Id { get; set; } = Guid.NewGuid();
|
|
public Guid ConversationId { get; set; }
|
|
public AgentConversation? Conversation { get; set; }
|
|
|
|
public AgentRole Role { get; set; }
|
|
|
|
/// <summary>
|
|
/// Position in the conversation, 0-based. Timestamps are not enough to order a
|
|
/// transcript: a fast turn can produce two messages inside the same tick.
|
|
/// </summary>
|
|
public int Sequence { get; set; }
|
|
|
|
/// <summary>The visible text of the turn.</summary>
|
|
public string Content { get; set; } = string.Empty;
|
|
|
|
/// <summary>
|
|
/// JSON array of <c>{ name, input, result }</c> objects describing tool calls made
|
|
/// during this turn. Null on user turns and on assistant turns that used no tools.
|
|
/// </summary>
|
|
public string? ToolCallsJson { get; set; }
|
|
|
|
public DateTimeOffset CreatedAt { get; set; } = DateTimeOffset.UtcNow;
|
|
}
|