 yetone
|
6bf1c3987f
|
routing: a group's rule can hold hours of the day (Discord: send DeepSeek's or GLM's peak-price hours to another provider) — a rule takes time: {from, to, days} in this computer's local time, HH:MM to HH:MM; hours that end before they begin run past midnight and belong to the day they begin on (Friday's 22:00–08:00 takes Saturday's early hours), days are optional, and the hours are checked as a turn begins, so a turn that began inside them finishes where it went. The rule reads "time 09:00–18:00 Mon–Fri" on the group and in the trace; the editor types the hours into one pill (从 … 到 …), notes past midnight, marks a time that isn't one and picks days like the agents (every day, weekdays, weekends, or each); magpie group rule add … time=14:00-18:00 days=mon-fri and the TUI's rule line take it too. Older versions ignore the field, and a rule with only hours then matches nothing there. TestRuleTimeMatches and TestRuleTimeWindowAtTurnStart (injected clock) and routing-time.test.cjs fail without the change. A sandbox magpie with two fake vendors sent a request inside the hours to the rule's member and one outside them, or on another day, to the first. go vet, go build -tags nogui (darwin, GOOS=linux), GOOS=windows go build, go test -tags nogui ./... pass; routing-time, routing-compact, routing-manual, routing-steady, settings-groups, click-scroll, old-webkit pass.
|
2026-10-01 15:04:16 +08:00 |
|
|
yetone
|
3fe2ff9958
|
rules: a routing group's rule can be for the agent compacting its conversation, so a cheaper, faster model writes the summary while the conversation stays on its own (X, MR.Guo: 希望 Magpie 能像 OpenCode 一样支持单独指定 context compression model：当 Agent 需要做上下文压缩时…指定一个更便宜、更快的小模型专门负责总结和压缩历史上下文) — a compaction went to whatever model the conversation was on, the priciest in the group as often as not, and a long session paid for each summary at that price. A rule's new condition, compact (provider.Rule.Compact, "compacting"; magpie group rule add <group> use=<model> compact, compact=yes|no; the GUI editor's 压缩上下文时 / compacting toggle; the TUI hint), matches a request that is an agent compacting, told from the agent's own prompt (internal/gateway/compacting.go, read from each one's release): the system prompt of Claude Code 2.1's /compact and auto-compact, OpenCode's compaction agent, Pi, Gemini CLI's state_snapshot and Qwen Code, or the last user message of Claude Code's, Codex's (and magpie's own for a Codex compaction_trigger) and Kimi Code's; a prompt only quoted earlier in a conversation doesn't count. Only in a group with such a rule is a request looked at this way, so nothing else changes. A compaction is decided on its own, never held by the turn's rule, its classifier never asked; it isn't remembered for the turn, and neither the conversation's affinity nor the turn's size is taken from its answer, so the requests after it go on on the model the turn was on and its cache (without that, the next request of a Claude Code turn followed the summary to the cheap model). A member whose context is less than the compaction is passed over for the next rule that matches, or the group's order (RuleHit.Small). The trace tells it ("The agent is compacting the conversation, so rule 1 sends the summary to …", the models passed over named), and the editor notes a compact rule's model taking less than another member. Docs: start#rules (en, zh). Tested: internal/gateway/compacting_test.go — each agent's prompt taken, a plain and a quoted one not; in a group with rules compact→a/small and tokens≥2000→b/big, a long Claude Code turn on b, its /compact to a (Compact, N=1), the turn's next request back on b held, Codex's compaction to a too; a 70k-token compaction skipping a (64k) for rule 2 with Small [a/small]; a group with no compact rule routing a compaction as before; ParseRule/Line/Conditions/Matches for compact. TestRuleCompact failed with the affinity kept from the summary. internal/gui/tests/routing-compact.test.cjs, Chromium and WebKit, en and zh: the compaction in the trace, the rule's words, the toggle on and off with nothing moved, the room note, a compacting-only rule saved as compact:true and not refused. routing-kind/served/signin/steady, click-scroll and old-webkit pass. go vet, go build -tags nogui (and GOOS=linux, GOOS=windows) and go test -tags nogui ./... pass. Not tried against real agents' compactions through the gateway (the prompts are matched as read from their releases) nor in the real Wails window.
|
2026-09-30 10:24:18 +08:00 |
|
|
yetone
|
ba9f0c610d
|
routing: a group's member may be a model at an effort of its own, provider/model:low (#189, @luw2007: model + effort 才是完整的组合项) — a member was only a model, reasoning as the agent asked or as the group's classifier picked, so "the same model, cheap and quick" and "the same model, thinking hard" could not both be in a group, nor could a rule send to one of them. A member now may carry an effort (none … max): a request routed to it is asked for that much reasoning whatever the agent asked or effort=auto picked, at the level the model has nearest, while the vendor is asked for the model's own id. The same model at two efforts is two members that rules (use=m:high), stickiness and the trace tell apart; an id whose own last part is a colon (a :free, :7b, :0) stays the model's. The group's reasoning levels are those of its members that follow the agent, and the editor has a reasoning button on each member. Tested with a fake provider through a sandbox magpie: the upstream bodies carry the fixed effort in Chat, Responses and Anthropic shapes, and the routing view shows m1:low / m1:high as two seats, light and dark
|
2026-09-29 13:45:29 +08:00 |
|
|
yetone
|
c286c218e3
|
routing: a group's intent classifier can be a routing group, failing over as any request to it does (lantern2077 on Discord: 意图分类模型能否也选择路由组？现在只能选择单个模型，缺乏回退保障) — the classifier was one model, and when it was down no intent matched. It can now be group/<id>, asked through magpie as before, so its first model failing the next is asked. A group can't be its own classifier, one that classifies for another can't be removed until that one chooses another, a rename carries over, and a classifier's own call asks no classifier in turn, so two groups classifying with each other don't loop. The GUI and TUI pickers list the other groups. Tested: the classifier group's first model 500s and the next answers and the rule matches; two groups set to classify with each other finish at once; checked in a sandbox magpie web that the picker lists Cls and not the group being edited, and via the CLI that self-classifying and removing a classifier group are refused
|
2026-09-29 10:25:41 +08:00 |
|
|
yetone
|
8630e4c62a
|
tui: the routing page (3, was groups) does what the app's Routing view does — an open group lists its rules under its models, one list: n adds a rule typed as magpie group rule add takes it (use=<model> tokens=200k images effort=… agents=… intent="…" classifier=<model> at=<n>), ↵ types one again in place, J/K moves it, d twice removes it; c picks the classifier, s how long a conversation stays, x the context agents are told, R renames the group (the page follows it); a found group that was removed is listed last and u brings it back; smart routing is shown as smart, not in order, and o goes round all four. The library page adds too: a adds an MCP server (a URL, or a command) or skills from a GitHub repository or a folder, e changes a server (its env, headers and agents kept), u fetches a skill from GitHub again, s syncs the library into the agents. The rule parsing moves to provider (ParseRule, RuleWords, Rule.Line) and a typed server to library (ServerOf), for the CLI and the TUI alike. Asked on Discord
|
2026-09-26 23:08:58 +08:00 |
|
|
yetone
|
d4135b55e1
|
group: a routing group's models can be another group — group/<id> as a member, routed in its place by its own routing and rules, so Coder can be deepseek-v4-pro then Fast, a Coder rule can send a turn to group/fast and Fast's own rules pick within it; a group can't be in itself directly or through another (refused on save, left out of the picker), groups go at most 8 deep, and one can't be removed while another has it. The Routing tab shows a group in a group as a heading with its models under it, the packet going through it, and the log says what each group's rules and classifier decided
|
2026-09-26 13:41:17 +08:00 |
|
|
yetone
|
bc720c8d87
|
group: a rule can be for a kind of message — intent="writing or fixing tests": as a user's turn begins, the group's classifier (any model magpie has, best a small fast one) is asked once which of the intents that may match the message is, and the turn stays with the answer through its tool rounds; it isn't asked when no intent rule could decide, the same message isn't asked about twice, and when it fails or is slow (8s) no intent matches and it rests 30s. The trace says what it was asked and said; Routing's rule editor has "asks for" and the classifier picker, and magpie group rule add … intent=… classifier=… / magpie group rule classifier from the terminal. A sub-millisecond reply no longer shows as NaN s in the trace
|
2026-09-26 11:09:13 +08:00 |
|
|
yetone
|
696c8e70e7
|
group: rules send a kind of turn to a chosen model — a group's Rules (Routing view, or magpie group rule add|rm|mv) put one of its models first for a turn that is long (tokens ≥ N), carries an image, asks for reasoning (on or at least a level) or comes from certain agents; the first that matches decides as the turn begins and the turn stays there through its tool rounds, moving only if it outgrows its model's context, the others staying behind it for failover; a subagent keeps a place of its own so a rule can send it elsewhere while its agent stays; the trace says which rule sent a turn where. Turn affinity also now sees past the system message Claude Code puts after tool results, which made each of its tool rounds look like a new turn
|
2026-09-26 03:53:33 +08:00 |
|