1
import json
2
from twelvelabs import TwelveLabs, IngestionConfig, EnrichmentConfig_Description, TextParam
3
from twelvelabs.types.text_param_format import TextParamFormat_JsonSchema
4
5
client = TwelveLabs(api_key="<YOUR_API_KEY>")
6
7
# Create a knowledge store with a domain-specific ingestion configuration
8
store = client.knowledge_stores.create(
9
    name="Microdrama Source Material",
10
    ingestion_config=IngestionConfig(
11
        enrichment_config=EnrichmentConfig_Description(
12
            description="Focus on characters and their emotions, interpersonal dynamics, conflicts, tension points, visual mood shifts, dialogue tone, and dramatic turning points. Track recurring characters across videos."
13
        )
14
    ),
15
)
16
store_id = store.id
17
18
# Before continuing, add video assets to your knowledge store and wait
19
# for processing to complete. See the Quickstart for details.
20
21
# Define a JSON schema for the narrative output
22
drama_schema = {
23
    "type": "object",
24
    "properties": {
25
        "title": {"type": "string"},                # Title for the micro-drama
26
        "logline": {"type": "string"},               # One-sentence summary
27
        "characters": {
28
            "type": "array",
29
            "items": {
30
                "type": "object",
31
                "properties": {
32
                    "name": {"type": "string"},      # Character name
33
                    "role": {"type": "string"},      # Protagonist, antagonist, etc.
34
                    "arc": {"type": "string"},       # Character development summary
35
                },
36
            },
37
        },
38
        "scenes": {
39
            "type": "array",
40
            "items": {
41
                "type": "object",
42
                "properties": {
43
                    "scene_number": {"type": "integer"},     # Position in sequence
44
                    "description": {"type": "string"},       # What happens in the scene
45
                    "video_reference": {"type": "string"},   # Source video identifier
46
                    "timestamp": {"type": "string"},         # Clip timestamp
47
                    "dramatic_function": {"type": "string"}, # Role in the narrative arc
48
                },
49
            },
50
        },
51
        "central_conflict": {"type": "string"},      # Main dramatic tension
52
        "resolution": {"type": "string"},            # How the conflict resolves
53
    },
54
}
55
56
# Request a narrative outline
57
response = client.responses.create(
58
    knowledge_store_id=store_id,
59
    instructions="You are a micro-drama creator. Analyze the video collection for narrative potential. Identify characters, conflicts, and dramatic moments. Construct a compelling 60-second micro-drama outline.",
60
    input=[{"type": "message", "role": "user", "content": "Create a micro-drama from these videos. Focus on the strongest emotional arc you can find."}],
61
    text=TextParam(format=TextParamFormat_JsonSchema(name="micro_drama", schema_=drama_schema)),
62
)
63
64
# Parse and read the outline
65
session_id = response.session_id
66
for output in response.output:
67
    if output.type == "message":
68
        for content in output.content:
69
            drama = json.loads(content.text)
70
            print(f"Title: {drama['title']}")
71
            print(f"Logline: {drama['logline']}")
72
            print(f"Conflict: {drama['central_conflict']}")
73
            for scene in drama["scenes"]:
74
                print(f"  Scene {scene['scene_number']}: {scene['description']}")
75
76
# Refine the output in a follow-up turn
77
response = client.responses.create(
78
    knowledge_store_id=store_id,
79
    session_id=session_id,
80
    input=[{"type": "message", "role": "user", "content": "Make the conflict sharper. Can you find a stronger turning point moment?"}],
81
    text=TextParam(format=TextParamFormat_JsonSchema(name="micro_drama", schema_=drama_schema)),
82
)
83
84
# Process the refined output
85
for output in response.output:
86
    if output.type == "message":
87
        for content in output.content:
88
            drama = json.loads(content.text)
89
            print(f"\nRefined Title: {drama['title']}")
90
            print(f"Logline: {drama['logline']}")
91
            print(f"Conflict: {drama['central_conflict']}")
92
            for scene in drama["scenes"]:
93
                print(f"  Scene {scene['scene_number']}: {scene['description']}")