@bicharts/chart-mcp 0.5.106 → 0.5.108
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +9 -2
- package/dist/index.mjs +65 -60
- package/package.json +3 -3
package/README.md
CHANGED
|
@@ -54,6 +54,7 @@ only render text); `structuredContent` is for code, so nothing has to string-par
|
|
|
54
54
|
| `geo` / `geoPoint` | region/basemap wiring, and how many points actually placed |
|
|
55
55
|
| `data` **or** `files` | the render payload inline, **or** the paths written — never both |
|
|
56
56
|
| `hostContract`, `build` | the contract the code targets, and the bundle that served the call |
|
|
57
|
+
| `recoveredAfterMs` | present only when the response stream was cut and the chart was fetched back by its correlation: milliseconds from the start of the call to delivery. `creditCost` is null then |
|
|
57
58
|
|
|
58
59
|
`data` carries the appended `__geoIso__` / `__geoLat__` / `__geoLon__` columns, which a caller
|
|
59
60
|
cannot reconstruct from the source data — so a geo chart needs either it or `files.data`.
|
|
@@ -134,6 +135,7 @@ Two ways to supply them — env wins over the file:
|
|
|
134
135
|
| `BIC_URL` | override backend (default `https://bizintelligencechampions.com`) |
|
|
135
136
|
| `BIC_MCP_REASONING` | OPTIONAL override of the default reasoning mode (else `""` = leave-to-visual, same as the PBI visual) |
|
|
136
137
|
| `BIC_MCP_MODEL` | OPTIONAL override of the default model (else `""` = the account's IsDefault model, same as the visual) |
|
|
138
|
+
| `BIC_MCP_TIMEOUT_MS` | OPTIONAL: how long one `generate_chart` call waits, recovery included (default `900000`, 15 minutes) |
|
|
137
139
|
|
|
138
140
|
**Defaults match the Power BI visual.** Unset, the MCP sends `model=""` (server
|
|
139
141
|
resolves the `IsDefault` model), `reasoning_mode=""` (leave-to-visual heuristic), and
|
|
@@ -156,8 +158,13 @@ needs**:
|
|
|
156
158
|
- Fallback for hosts that don't reset on progress: raise the host's per-tool
|
|
157
159
|
timeout (Claude Code: `MCP_TOOL_TIMEOUT` in ms, set in the environment Claude
|
|
158
160
|
Code itself runs in — e.g. the `env` block of `settings.json`).
|
|
159
|
-
-
|
|
160
|
-
|
|
161
|
+
- **Built in:** when the response stream is cut *after* the server accepted the
|
|
162
|
+
request, the MCP does not give up or re-generate. Inside the same tool call it
|
|
163
|
+
checks for the finished chart by its `correlationId` (a fetch that can never call a
|
|
164
|
+
model), backing off between checks, until the chart arrives or the recovery window
|
|
165
|
+
or `BIC_MCP_TIMEOUT_MS` runs out, whichever is sooner. A recovered chart carries
|
|
166
|
+
`recoveredAfterMs`; when nothing arrives, the error keeps the `correlationId` to
|
|
167
|
+
quote to support.
|
|
161
168
|
|
|
162
169
|
`BIC_MCP_REASONING=CP`/`1P` *would* finish faster, but that trades away visual parity
|
|
163
170
|
(no validator) — prefer raising the timeout.
|