PromptSnapshot

PromptSnapshot

The complete prompt configuration captured at deployment time. Contains model config, messages, tools, and variable definitions.

Overview

PromptSnapshot is the prompt payload inside a Deployment. When you call get_deployment() or get_latest_deployment(), the returned Deployment.prompt field is a PromptSnapshot.

from adaline_api.models.prompt_snapshot import PromptSnapshot

---

Fields

config · PromptSnapshotConfig · required

Model provider and settings. See PromptSnapshotConfig.

messages · list[PromptMessage] · required

Array of prompt messages (role and content). See PromptMessage.

tools · list[ToolFunction] · required

Array of tool/function definitions available to the model. See ToolFunction.

variables · list[PromptVariable] · required

Array of variable definitions used in the prompt template. See PromptVariable.

---

Examples

Accessing from a Deployment

from adaline.main import Adaline

adaline = Adaline()

deployment = await adaline.get_latest_deployment(
    prompt_id="prompt_abc123",
    deployment_environment_id="environment_abc123"
)

prompt: PromptSnapshot = deployment.prompt

print(f"Provider: {prompt.config.provider_name}")
print(f"Model: {prompt.config.model}")
print(f"Messages: {len(prompt.messages)}")
print(f"Tools: {len(prompt.tools)}")
print(f"Variables: {[v.name for v in prompt.variables]}")

Inspecting Messages and Tools

from adaline_api.models.prompt_message import PromptMessage
from adaline_api.models.text_content import TextContent

prompt = deployment.prompt

for msg in prompt.messages:
    for c in msg.content:
        if isinstance(c.actual_instance, TextContent):
            print(f"[{msg.role}] {c.actual_instance.value}")

for tool in prompt.tools:
    fn = tool.definition.var_schema
    print(f"Tool: {fn.name} — {fn.description}")

Using with a Provider SDK

import re
from adaline.main import Adaline
from openai import OpenAI

adaline = Adaline()
openai_client = OpenAI()

deployment = await adaline.get_latest_deployment(
    prompt_id="prompt_abc123",
    deployment_environment_id="environment_abc123",
)

# Replace {{variable_name}} placeholders with your runtime values.
# See PromptVariable for the full pattern (text + image + PDF).
variables = {"user_name": "Alice"}
pattern = re.compile(r"\{\{(\w+)\}\}")
for message in deployment.prompt.messages:
    for c in message.content:
        if c.modality == "text":
            c.value = pattern.sub(
                lambda m: str(variables.get(m.group(1), m.group(0))),
                c.value,
            )

config = deployment.prompt.config
response = openai_client.chat.completions.create(
    model=config.model,
    messages=[...],  # Build from deployment.prompt.messages
    temperature=config.settings.get("temperature"),
    max_tokens=config.settings.get("maxTokens"),
)

---

Serialization

from adaline_api.models.prompt_snapshot import PromptSnapshot

prompt = deployment.prompt

d = prompt.to_dict()
j = prompt.to_json()

restored = PromptSnapshot.from_dict(d)
restored = PromptSnapshot.from_json(j)

---

JSON Example

{
  "config": {
    "provider_name": "openai",
    "provider_id": "provider_abc123",
    "model": "gpt-4o",
    "settings": { "temperature": 0.7, "maxTokens": 1000 }
  },
  "messages": [
    {
      "role": "system",
      "content": [
        { "modality": "text", "value": "You are a helpful assistant." }
      ]
    }
  ],
  "tools": [],
  "variables": [
    { "name": "user_name", "modality": "text" }
  ]
}