File size: 10,339 Bytes
193db9d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
import json
from unittest.mock import patch

import pytest

from workflows.errors import CyclicDependencyError, WorkflowError
from workflows.executors import (
    create_processed_inputs,
    execute_model_step,
    execute_workflow,
    lower,
    upper,
)
from workflows.structs import InputField, ModelStep, OutputField, Workflow

# Tests for utility functions


def test_upper():
    """Test the upper function with different input types."""
    assert upper("hello") == "HELLO"
    assert upper("Hello World") == "HELLO WORLD"
    assert upper("") == ""
    # Non-string inputs should be returned unchanged
    assert upper(123) == 123
    assert upper([1, 2, 3]) == [1, 2, 3]
    assert upper(None) is None


def test_lower():
    """Test the lower function with different input types."""
    assert lower("HELLO") == "hello"
    assert lower("Hello World") == "hello world"
    assert lower("") == ""
    # Non-string inputs should be returned unchanged
    assert lower(123) == 123
    assert lower([1, 2, 3]) == [1, 2, 3]
    assert lower(None) is None


# Tests for create_processed_inputs


def test_create_processed_inputs_basic():
    """Test basic input processing without transformations."""
    step = ModelStep(
        id="test_step",
        model="gpt-4",
        provider="openai",
        call_type="llm",
        system_prompt="Test prompt",
        input_fields=[InputField(name="text", description="Input text", variable="input_text")],
        output_fields=[],
    )
    available_vars = {"input_text": "Hello World"}

    result = create_processed_inputs(step, available_vars)
    assert result == {"text": "Hello World"}


def test_create_processed_inputs_with_transformation():
    """Test input processing with transformation functions."""
    step = ModelStep(
        id="test_step",
        model="gpt-4",
        provider="openai",
        call_type="llm",
        system_prompt="Test prompt",
        input_fields=[
            InputField(name="upper_text", description="Uppercase text", variable="input_text", func="upper"),
            InputField(name="lower_text", description="Lowercase text", variable="input_caps", func="lower"),
        ],
        output_fields=[],
    )
    available_vars = {"input_text": "hello", "input_caps": "WORLD"}

    result = create_processed_inputs(step, available_vars)
    assert result == {"upper_text": "HELLO", "lower_text": "world"}


def test_create_processed_inputs_missing_var():
    """Test that appropriate error is raised when a variable is missing."""
    step = ModelStep(
        id="test_step",
        model="gpt-4",
        provider="openai",
        call_type="llm",
        system_prompt="Test prompt",
        input_fields=[InputField(name="text", description="Input text", variable="missing_var")],
        output_fields=[],
    )
    available_vars = {"input_text": "Hello World"}

    with pytest.raises(KeyError):
        create_processed_inputs(step, available_vars)


def test_create_processed_inputs_unknown_func():
    """Test that appropriate error is raised when an unknown function is specified."""
    step = ModelStep(
        id="test_step",
        model="gpt-4",
        provider="openai",
        call_type="llm",
        system_prompt="Test prompt",
        input_fields=[InputField(name="text", description="Input text", variable="input_text", func="unknown_func")],
        output_fields=[],
    )
    available_vars = {"input_text": "Hello World"}

    # This should raise an error when the function isn't found
    with pytest.raises(Exception):
        create_processed_inputs(step, available_vars)


# Tests for execute_model_step


@patch("workflows.executors.litellm.completion")
def test_execute_model_step_success(mock_completion):
    """Test successful execution of a model step with mocked litellm response."""
    # Mock the litellm response
    mock_response = {"choices": [{"message": {"content": json.dumps({"summary": "This is a summary"})}}]}
    mock_completion.return_value = mock_response

    # Create a test step
    step = ModelStep(
        id="summarize",
        model="gpt-3.5-turbo",
        provider="openai",
        call_type="llm",
        system_prompt="Summarize the text",
        input_fields=[InputField(name="text", description="Text to summarize", variable="input_text")],
        output_fields=[OutputField(name="summary", description="Summary of the text", type="str")],
    )

    # Execute the step
    result = execute_model_step(step, {"input_text": "Long text to be summarized..."})

    # Verify the results
    assert result == {"summary": "This is a summary"}

    # Verify the litellm call was made correctly
    mock_completion.assert_called_once()
    args, kwargs = mock_completion.call_args
    assert kwargs["model"] == "gpt-3.5-turbo"
    assert "Summarize the text" in kwargs["messages"][0]["content"]


@patch("workflows.executors.litellm.completion")
def test_execute_model_step_error(mock_completion):
    """Test handling of errors in model step execution."""
    # Make litellm raise an exception
    mock_completion.side_effect = Exception("API Error")

    # Create a test step
    step = ModelStep(
        id="summarize",
        model="gpt-3.5-turbo",
        provider="openai",
        call_type="llm",
        system_prompt="Summarize the text",
        input_fields=[InputField(name="text", description="Text to summarize", variable="input_text")],
        output_fields=[OutputField(name="summary", description="Summary of the text", type="str")],
    )

    # Execute the step - should raise an exception
    with pytest.raises(Exception):
        execute_model_step(step, {"input_text": "Long text to be summarized..."})


# Tests for execute_workflow


@patch("workflows.executors.execute_model_step")
def test_execute_workflow_simple(mock_execute_step):
    """Test execution of a simple workflow with a single step."""
    # Configure mock to return expected outputs
    mock_execute_step.return_value = {"summary": "This is a summary"}

    # Create a simple workflow
    step = ModelStep(
        id="summarize",
        model="gpt-3.5-turbo",
        provider="openai",
        call_type="llm",
        system_prompt="Summarize the text",
        input_fields=[InputField(name="text", description="Text to summarize", variable="input_text")],
        output_fields=[OutputField(name="summary", description="Summary of the text", type="str")],
    )

    workflow = Workflow(steps={"summarize": step}, inputs=["input_text"], outputs={"summary": "summarize.summary"})

    # Execute the workflow
    result = execute_workflow(workflow, {"input_text": "Long text to be summarized..."})

    # Verify the results
    assert result == {"summary": "This is a summary"}

    # Verify execute_model_step was called correctly
    mock_execute_step.assert_called_once()


@patch("workflows.executors.execute_model_step")
def test_execute_workflow_multi_step(mock_execute_step):
    """Test execution of a multi-step workflow with dependencies."""

    # Configure mock to return different values based on the step
    def side_effect(step, available_vars):
        if step.id == "extract":
            return {"entities": ["Apple", "product"]}
        elif step.id == "analyze":
            return {"sentiment": "positive"}
        return {}

    mock_execute_step.side_effect = side_effect

    # Create extract step
    extract_step = ModelStep(
        id="extract",
        model="gpt-3.5-turbo",
        provider="openai",
        call_type="llm",
        system_prompt="Extract entities",
        input_fields=[InputField(name="text", description="Text to analyze", variable="input_text")],
        output_fields=[OutputField(name="entities", description="Extracted entities", type="list[str]")],
    )

    # Create analyze step that depends on extract step
    analyze_step = ModelStep(
        id="analyze",
        model="gpt-4",
        provider="openai",
        call_type="llm",
        system_prompt="Analyze sentiment",
        input_fields=[InputField(name="entities", description="Entities to analyze", variable="extract.entities")],
        output_fields=[OutputField(name="sentiment", description="Sentiment analysis", type="str")],
    )

    workflow = Workflow(
        steps={"extract": extract_step, "analyze": analyze_step},
        inputs=["input_text"],
        outputs={"entities": "extract.entities", "sentiment": "analyze.sentiment"},
    )

    # Execute the workflow
    result = execute_workflow(workflow, {"input_text": "Apple is launching a new product tomorrow."})

    # Verify the results
    assert result == {"entities": ["Apple", "product"], "sentiment": "positive"}

    # Verify execute_model_step was called twice (once for each step)
    assert mock_execute_step.call_count == 2


def test_execute_workflow_missing_input():
    """Test that an error is raised when a required input is missing."""
    step = ModelStep(
        id="summarize",
        model="gpt-3.5-turbo",
        provider="openai",
        call_type="llm",
        system_prompt="Summarize the text",
        input_fields=[InputField(name="text", description="Text to summarize", variable="input_text")],
        output_fields=[OutputField(name="summary", description="Summary of the text", type="str")],
    )

    workflow = Workflow(steps={"summarize": step}, inputs=["input_text"], outputs={"summary": "summarize.summary"})

    # Execute with missing input
    with pytest.raises(WorkflowError, match="Missing required workflow input"):
        execute_workflow(workflow, {})


@patch("workflows.executors.create_dependency_graph")
def test_execute_workflow_cyclic_dependency(mock_dependency_graph):
    """Test that a cyclic dependency in the workflow raises an appropriate error."""
    # Make create_dependency_graph raise a CyclicDependencyError
    mock_dependency_graph.side_effect = CyclicDependencyError()

    step = ModelStep(
        id="test",
        model="gpt-3.5-turbo",
        provider="openai",
        call_type="llm",
        system_prompt="Test",
        input_fields=[],
        output_fields=[],
    )

    workflow = Workflow(steps={"test": step}, inputs=[], outputs=[])

    # This should propagate the CyclicDependencyError
    with pytest.raises(CyclicDependencyError):
        execute_workflow(workflow, {})