File size: 4,630 Bytes
ed6bec6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
# generator/templates_context.py

import random
from typing import List, Dict, Any

from .meaning_model import StructuredMeaning, Context
from .encoder import encode_term


def generate_context_rich_samples(
    dictionaries: Dict[str, Any],
    count: int,
    source_language: str = "en",
) -> List[Dict[str, Any]]:
    """
    Generates samples that combine:
    - actor
    - action
    - object
    - time context
    - place context
    - activity context
    """

    actions = dictionaries.get("actions", [])
    objects = dictionaries.get("objects", [])
    actors = dictionaries.get("actors", [])
    times = dictionaries.get("context_time", [])
    places = dictionaries.get("context_place", [])
    activities = dictionaries.get("context_activity", [])

    results = []

    if not actions or not objects or not actors:
        raise ValueError("Missing required dictionaries: actions, objects, actors.")

    for _ in range(count):
        action = random.choice(actions)
        obj = random.choice(objects)
        actor = random.choice(actors)

        time_ctx = random.choice(times) if times else None
        place_ctx = random.choice(places) if places else None
        activity_ctx = random.choice(activities) if activities else None

        action_id = action.get("id", "action.unknown")
        object_id = obj.get("id", "object.unknown")
        actor_id = actor.get("id", "actor.unknown")

        time_id = time_ctx.get("id") if time_ctx else None
        place_id = place_ctx.get("id") if place_ctx else None
        activity_id = activity_ctx.get("id") if activity_ctx else None

        input_text = f"{actor_id} {action_id} the {object_id}"
        if place_id:
            input_text += f" in the {place_id}"
        if activity_id:
            input_text += f" while {activity_id}"
        if time_id:
            input_text += f" ({time_id})"
        input_text += "."

        meaning = StructuredMeaning(
            actor=actor_id,
            action=action_id,
            object=object_id,
            modifiers=[],
            emotion=None,
            context=Context(
                time=time_id,
                place=place_id,
                activity=activity_id,
            ),
            intent="intent.describe_event",
            meta={
                "source_language": source_language,
                "confidence": 0.95,
            },
        )

        actor_enc = encode_term(actor_id, dictionaries)
        action_enc = encode_term(action_id, dictionaries)
        object_enc = encode_term(object_id, dictionaries)
        time_enc = encode_term(time_id, dictionaries) if time_id else None
        place_enc = encode_term(place_id, dictionaries) if place_id else None
        activity_enc = encode_term(activity_id, dictionaries) if activity_id else None

        glyphic_human = (
            f"ACTOR[{actor_enc['human']}] "
            f"ACTION[{action_enc['human']}] "
            f"OBJECT[{object_enc['human']}]"
        )
        if time_enc:
            glyphic_human += f" CONTEXT.TIME[{time_enc['human']}]"
        if place_enc:
            glyphic_human += f" CONTEXT.PLACE[{place_enc['human']}]"
        if activity_enc:
            glyphic_human += f" CONTEXT.ACTIVITY[{activity_enc['human']}]"
        glyphic_human += " INTENT.describe"

        glyphic_compact = (
            f"ACT{{{actor_enc['compact']}}} "
            f"ACTN{{{action_enc['compact']}}} "
            f"OBJ{{{object_enc['compact']}}}"
        )
        if time_enc:
            glyphic_compact += f" CTX{{TIME:{time_enc['compact']}}}"
        if place_enc:
            glyphic_compact += f" CTX{{PLACE:{place_enc['compact']}}}"
        if activity_enc:
            glyphic_compact += f" CTX{{ACT:{activity_enc['compact']}}}"
        glyphic_compact += " INT{DESC}"

        glyphic_tokens = (
            f"{actor_enc['tokens']} "
            f"{action_enc['tokens']} "
            f"{object_enc['tokens']}"
        )
        if time_enc:
            glyphic_tokens += f" <CTX:TIME:{time_enc['compact']}>"
        if place_enc:
            glyphic_tokens += f" <CTX:PLACE:{place_enc['compact']}>"
        if activity_enc:
            glyphic_tokens += f" <CTX:ACT:{activity_enc['compact']}>"
        glyphic_tokens += " <INT:DESC>"

        results.append(
            {
                "input_text": input_text,
                "glyphic_output_human": glyphic_human,
                "glyphic_output_compact": glyphic_compact,
                "glyphic_output_tokens": glyphic_tokens,
                "structured_meaning": meaning.to_dict(),
            }
        )

    return results