Spaces:
Sleeping
Sleeping
File size: 5,673 Bytes
83fe4f9 fb38df2 83fe4f9 fb38df2 83fe4f9 fb38df2 83fe4f9 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 | import json
from backend.ai.client import llm_client, render_prompt
from backend.ai.prompts import BUILD_WORKFLOW_PROMPT
from backend.models.schemas import BuildWorkflowRequest, WorkflowDefinition, WorkflowStep
def build_workflow_definition(request: BuildWorkflowRequest, owner: dict) -> dict:
if llm_client.is_ready():
prompt = render_prompt(
BUILD_WORKFLOW_PROMPT,
selected_option=json.dumps(request.selected_option, indent=2),
task_name=request.task_name,
owner_email=owner.get("email", "owner@example.com"),
spreadsheet_id=owner.get("spreadsheet_id", ""),
inventory_sheet=owner.get("spreadsheet_config", {}).get("inventory_sheet", "Inventory"),
orders_sheet=owner.get("spreadsheet_config", {}).get("orders_sheet", "Orders"),
item_column="item",
stock_column="stock",
uploaded_data_summary=json.dumps(owner.get("uploaded_data_summary", []), indent=2),
)
return llm_client.generate_json(prompt)
trigger_type = "schedule" if "summary" in request.task_name.lower() else "email_received"
steps: list[WorkflowStep] = []
if trigger_type == "email_received":
steps = [
WorkflowStep(
id="step_extract",
action="ai_extract",
params={
"input": "{{trigger.email.body}}",
"extract_fields": ["customer_name", "items", "pickup_date", "needs_clarification"],
"output_var": "order_data",
},
on_error="notify_owner",
),
WorkflowStep(
id="step_read_inventory",
action="sheets_read",
params={
"spreadsheet_id": "{{config.spreadsheet_id}}",
"sheet_name": "Inventory",
"output_var": "inventory_rows",
},
on_error="abort",
),
WorkflowStep(
id="step_check_stock",
action="condition",
params={
"check": "not {{step_extract.order_data.needs_clarification}}",
"if_true": "step_write_order",
"if_false": "step_notify_ambiguity",
},
on_error="abort",
),
WorkflowStep(
id="step_write_order",
action="sheets_write",
params={
"spreadsheet_id": "{{config.spreadsheet_id}}",
"sheet_name": "Orders",
"data": {
"customer": "{{step_extract.order_data.customer_name}}",
"items": "{{step_extract.order_data.items}}",
"priority": "high" if "restaurant" in request.task_name.lower() else "normal",
},
},
on_error="notify_owner",
),
WorkflowStep(
id="step_compose_reply",
action="ai_compose",
params={
"context": "{{step_extract.order_data}}",
"tone": owner.get("preferred_tone", "friendly"),
"output_var": "reply_body",
},
on_error="notify_owner",
),
WorkflowStep(
id="step_send_email",
action="send_email",
params={
"to": "{{trigger.email.from}}",
"subject": "Order confirmation",
"body": "{{step_compose_reply.reply_body}}",
},
on_error="notify_owner",
),
WorkflowStep(
id="step_notify_ambiguity",
action="notify_owner",
params={
"message": "This order needs clarification before FlowPilot can continue.",
"severity": "warning",
"options": ["Reply myself", "Ask customer for details"],
},
on_error="skip",
),
]
else:
steps = [
WorkflowStep(
id="step_collect_orders",
action="sheets_read",
params={
"spreadsheet_id": "{{config.spreadsheet_id}}",
"sheet_name": "Orders",
"output_var": "order_rows",
},
on_error="abort",
),
WorkflowStep(
id="step_compose_summary",
action="ai_compose",
params={
"context": "{{step_collect_orders.order_rows}}",
"tone": "helpful",
"output_var": "summary_body",
},
on_error="notify_owner",
),
WorkflowStep(
id="step_send_summary",
action="send_email",
params={
"to": owner.get("email", "owner@example.com"),
"subject": "Weekly order summary",
"body": "{{step_compose_summary.summary_body}}",
},
on_error="notify_owner",
),
]
workflow = WorkflowDefinition(
name=request.task_name,
description=request.selected_option.get("name", request.task_description),
trigger={"type": trigger_type, "config": {"cron": "0 8 * * FRI"} if trigger_type == "schedule" else {}},
steps=steps,
)
return workflow.model_dump()
|