gaik 0.2.7__py3-none-any.whl → 0.2.9__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (5) hide show
  1. {gaik-0.2.7.dist-info → gaik-0.2.9.dist-info}/METADATA +68 -40
  2. {gaik-0.2.7.dist-info → gaik-0.2.9.dist-info}/RECORD +5 -5
  3. {gaik-0.2.7.dist-info → gaik-0.2.9.dist-info}/WHEEL +0 -0
  4. {gaik-0.2.7.dist-info → gaik-0.2.9.dist-info}/licenses/LICENSE +0 -0
  5. {gaik-0.2.7.dist-info → gaik-0.2.9.dist-info}/top_level.txt +0 -0
{gaik-0.2.7.dist-info → gaik-0.2.9.dist-info}/METADATA RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: gaik
3
- Version: 0.2.7
3
+ Version: 0.2.9
4
4
  Summary: General AI Kit - Reusable AI/ML components for Python
5
5
  Author: GAIK Project
6
6
  License: MIT License
@@ -98,58 +98,65 @@ pip install gaik[vision]
98
98
 
99
99
  ## Quick Start
100
100
 
101
- ### 1. Set up your provider API key
101
+ ### Extract Data from Text
102
102
 
103
- **OpenAI (default):**
103
+ Set your API key (choose one):
104
104
 
105
105
  ```bash
106
- export OPENAI_API_KEY='sk-...' # Get from: https://platform.openai.com/api-keys
106
+ export OPENAI_API_KEY='sk-...' # OpenAI (default)
107
+ export ANTHROPIC_API_KEY='sk-ant-...' # Anthropic
108
+ export GOOGLE_API_KEY='...' # Google
109
+ export AZURE_API_KEY='...' # Azure
110
+ export AZURE_ENDPOINT='https://...' # Azure (also required)
107
111
  ```
108
112
 
109
- **Anthropic:**
113
+ Then extract:
110
114
 
111
- ```bash
112
- export ANTHROPIC_API_KEY='sk-ant-...' # Get from: https://console.anthropic.com
113
- ```
115
+ ```python
116
+ from gaik.extract import SchemaExtractor
114
117
 
115
- **Google:**
118
+ # Using default OpenAI provider
119
+ extractor = SchemaExtractor("Extract name and age from text")
120
+ result = extractor.extract_one("Alice is 25 years old")
121
+ print(result)
122
+ # {'name': 'Alice', 'age': 25}
116
123
 
117
- ```bash
118
- export GOOGLE_API_KEY='...' # Get from: https://ai.google.dev
124
+ # Switch provider
125
+ extractor = SchemaExtractor("Extract name and age", provider="anthropic") # or "google", "azure"
119
126
  ```
120
127
 
121
- **Azure OpenAI:**
128
+ ### Convert PDF to Markdown
122
129
 
130
+ Requires: `pip install gaik[vision]`
131
+
132
+ Set environment variables first:
123
133
  ```bash
134
+ # For Azure OpenAI
124
135
  export AZURE_API_KEY='...'
125
136
  export AZURE_ENDPOINT='https://your-resource.openai.azure.com/'
126
- ```
137
+ export AZURE_DEPLOYMENT='gpt-4o' # Your deployment name
127
138
 
128
- ### 2. Simple Extraction
139
+ # OR for OpenAI
140
+ export OPENAI_API_KEY='sk-...'
141
+ ```
129
142
 
143
+ Then convert:
130
144
  ```python
131
- from gaik.extract import SchemaExtractor
145
+ from gaik.parsers import VisionParser, get_openai_config
132
146
 
133
- # Using default OpenAI provider
134
- extractor = SchemaExtractor("Extract name and age from text")
135
- result = extractor.extract_one("Alice is 25 years old")
136
- print(result)
137
- # {'name': 'Alice', 'age': 25}
147
+ # Configure
148
+ config = get_openai_config(use_azure=True) # or use_azure=False for OpenAI
149
+ parser = VisionParser(config)
138
150
 
139
- # Switch to Anthropic Claude
140
- extractor = SchemaExtractor(
141
- "Extract name and age from text",
142
- provider="anthropic"
143
- )
151
+ # Convert PDF (returns list of pages as Markdown strings)
152
+ pages = parser.convert_pdf("invoice.pdf", clean_output=True)
144
153
 
145
- # Use Google Gemini
146
- extractor = SchemaExtractor(
147
- "Extract name and age from text",
148
- provider="google"
149
- )
154
+ # Combine into single document
155
+ markdown = "\n\n".join(pages)
156
+ print(markdown)
150
157
  ```
151
158
 
152
- ### 3. Batch Extraction
159
+ ### Batch Extraction
153
160
 
154
161
  ```python
155
162
  from gaik.extract import dynamic_extraction_workflow
@@ -232,6 +239,8 @@ schema = InvoiceModel.model_json_schema()
232
239
 
233
240
  ## API Reference
234
241
 
242
+ ### Extraction API
243
+
235
244
  | Function/Class | Purpose |
236
245
  | ------------------------------- | ------------------------------------------------- |
237
246
  | `SchemaExtractor` | Reusable extractor with provider selection |
@@ -240,7 +249,15 @@ schema = InvoiceModel.model_json_schema()
240
249
  | `FieldSpec` | Define a single extraction field |
241
250
  | `ExtractionRequirements` | Collection of field specifications |
242
251
 
243
- ### Provider Parameters
252
+ ### Vision Parser API
253
+
254
+ | Function/Class | Purpose |
255
+ | --------------------- | ------------------------------------------ |
256
+ | `VisionParser` | PDF to Markdown converter using vision LLM |
257
+ | `get_openai_config()` | Helper to configure OpenAI/Azure API |
258
+ | `OpenAIConfig` | Configuration dataclass for vision parser |
259
+
260
+ ### Extraction Parameters
244
261
 
245
262
  ```python
246
263
  SchemaExtractor(
@@ -253,18 +270,29 @@ SchemaExtractor(
253
270
  )
254
271
  ```
255
272
 
256
- **Note:**
273
+ ### Vision Parser Parameters
274
+
275
+ ```python
276
+ VisionParser(config: OpenAIConfig)
277
+
278
+ get_openai_config(
279
+ use_azure: bool = True, # True for Azure, False for OpenAI
280
+ ) -> OpenAIConfig
281
+ ```
282
+
283
+ **Environment variables (auto-detected):**
257
284
 
258
- - IDEs with type checking (VS Code, PyCharm) will show autocomplete for `provider` parameter
259
- - Either `user_description` or `requirements` must be provided
260
- - Using `requirements` skips LLM parsing step (faster & cheaper)
285
+ - OpenAI: `OPENAI_API_KEY`
286
+ - Azure: `AZURE_API_KEY` + `AZURE_ENDPOINT` + `AZURE_DEPLOYMENT` (optional: `AZURE_API_VERSION`)
261
287
 
262
288
  ## Default Models
263
289
 
264
- - OpenAI: `gpt-4.1`
265
- - Anthropic: `claude-sonnet-4-5-20250929`
266
- - Google: `gemini-2.5-flash`
267
- - Azure: `gpt-4.1`
290
+ | Provider | Default Model |
291
+ | --------- | ---------------------------- |
292
+ | OpenAI | `gpt-4.1` |
293
+ | Anthropic | `claude-sonnet-4-5-20250929` |
294
+ | Google | `gemini-2.5-flash` |
295
+ | Azure | `gpt-4.1` (or your deployment) |
268
296
 
269
297
  ## Resources
270
298
 
{gaik-0.2.7.dist-info → gaik-0.2.9.dist-info}/RECORD RENAMED
@@ -11,8 +11,8 @@ gaik/providers/azure.py,sha256=lRCn04v57BM0vDHNtb9Uc3lzumrL4p3J0VuWB8RS5lY,2161
11
11
  gaik/providers/base.py,sha256=upH3S0QbKQSEC99mOF48LtKFcK5g3TFyd0GVxVuvgUk,1885
12
12
  gaik/providers/google.py,sha256=t5p4uBb3v81KgZ-JQ7N7oKIS-0doCem1IV4VThfLKIY,1690
13
13
  gaik/providers/openai.py,sha256=OUyc_a01oNEqtMO8KnHZmwqJp_hEQQC1FXfXelNNGPU,1559
14
- gaik-0.2.7.dist-info/licenses/LICENSE,sha256=jE12SAuBnlcHf16GXD5xfZNtSH6ds30U5LniIt_gRis,1087
15
- gaik-0.2.7.dist-info/METADATA,sha256=6lYKHurKCQK8FVemsdxfKhZ8OTRjpjb9_c70M_UT9AU,9048
16
- gaik-0.2.7.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
17
- gaik-0.2.7.dist-info/top_level.txt,sha256=p-eYZovskBEyr7C7GfsXIYQta9-fGyUSaFYDTX8oky8,5
18
- gaik-0.2.7.dist-info/RECORD,,
14
+ gaik-0.2.9.dist-info/licenses/LICENSE,sha256=jE12SAuBnlcHf16GXD5xfZNtSH6ds30U5LniIt_gRis,1087
15
+ gaik-0.2.9.dist-info/METADATA,sha256=Xz95byGbtzyZ5BsYalpFMJysGHqpA3cYuLuhhRBQdYY,10216
16
+ gaik-0.2.9.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
17
+ gaik-0.2.9.dist-info/top_level.txt,sha256=p-eYZovskBEyr7C7GfsXIYQta9-fGyUSaFYDTX8oky8,5
18
+ gaik-0.2.9.dist-info/RECORD,,
{gaik-0.2.7.dist-info → gaik-0.2.9.dist-info}/WHEEL RENAMED
File without changes