Skip to content

Commit fe7c262

Browse files
SergiiShcherbakSergiiShcherbak
committed
Docs and README
1 parent 7026b89 commit fe7c262

File tree

2 files changed

+84
-84
lines changed

2 files changed

+84
-84
lines changed

‎README.md‎

Lines changed: 42 additions & 42 deletions
Original file line numberDiff line numberDiff line change
@@ -39,158 +39,158 @@ Read more on the project [motivation](https://contextgem.dev/motivation.html) in
3939
<tr>
4040
<td>⭐</td>
4141
<td>
42-
<strong>Automated dynamic prompts</strong>:
43-
<small style="opacity: 0.8;">
42+
<strong>Automated dynamic prompts</strong><br>
43+
<sub>
4444
Automatically generates comprehensive prompts for your specific extraction needs.
45-
</small>
45+
</sub>
4646
</td>
4747
<td>✅</td>
4848
<td>❌</td>
4949
</tr>
5050
<tr>
5151
<td>⭐</td>
5252
<td>
53-
<strong>Automated data modelling and validators</strong>:
54-
<small style="opacity: 0.8;">
53+
<strong>Automated data modelling and validators</strong><br>
54+
<sub>
5555
Automatically creates data models and validation logic.
56-
</small>
56+
</sub>
5757
</td>
5858
<td>✅</td>
5959
<td>❌</td>
6060
</tr>
6161
<tr>
6262
<td>⭐</td>
6363
<td>
64-
<strong>Single, unified extraction pipeline (declarative, reusable, fully serializable)</strong>:
65-
<small style="opacity: 0.8;">
64+
<strong>Single, unified extraction pipeline (declarative, reusable, fully serializable)</strong><br>
65+
<sub>
6666
Allows to define a complete extraction workflow in a single, unified, reusable pipeline.
67-
</small>
67+
</sub>
6868
</td>
6969
<td>✅</td>
7070
<td>❌</td>
7171
</tr>
7272
<tr>
7373
<td>⭐</td>
7474
<td>
75-
<strong>Precise granular reference mapping (paragraphs & sentences)</strong>:
76-
<small style="opacity: 0.8;">
75+
<strong>Precise granular reference mapping (paragraphs & sentences)</strong><br>
76+
<sub>
7777
Automatically maps extracted data to the relevant parts of the document, which will always match
7878
in the source document, with customizable granularity.
79-
</small>
79+
</sub>
8080
</td>
8181
<td>✅</td>
8282
<td>❌</td>
8383
</tr>
8484
<tr>
8585
<td>⭐</td>
8686
<td>
87-
<strong>Justifications (reasoning backing the extraction)</strong>:
88-
<small style="opacity: 0.8;">
87+
<strong>Justifications (reasoning backing the extraction)</strong><br>
88+
<sub>
8989
Automatically provides justifications for each extraction, with
9090
customizable granularity.
91-
</small>
91+
</sub>
9292
</td>
9393
<td>✅</td>
9494
<td>❌</td>
9595
</tr>
9696
<tr>
9797
<td>⭐</td>
9898
<td>
99-
<strong>Neural segmentation (SaT)</strong>:
100-
<small style="opacity: 0.8;">
99+
<strong>Neural segmentation (SaT)</strong><br>
100+
<sub>
101101
Automatically segments the document into paragraphs and sentences using state-of-the-art SaT
102102
models, compatible with many languages.
103-
</small>
103+
</sub>
104104
</td>
105105
<td>✅</td>
106106
<td>❌</td>
107107
</tr>
108108
<tr>
109109
<td>⭐</td>
110110
<td>
111-
<strong>Multilingual support (I/O without prompting)</strong>:
112-
<small style="opacity: 0.8;">
111+
<strong>Multilingual support (I/O without prompting)</strong><br>
112+
<sub>
113113
Supports multiple languages in input and output without additional prompting.
114-
</small>
114+
</sub>
115115
</td>
116116
<td>✅</td>
117117
<td>❌</td>
118118
</tr>
119119
<tr>
120120
<td>⭐</td>
121121
<td>
122-
<strong>Grouped LLMs with role-specific tasks</strong>:
123-
<small style="opacity: 0.8;">
122+
<strong>Grouped LLMs with role-specific tasks</strong><br>
123+
<sub>
124124
Allows to easily group LLMs with different roles to process role-specific
125125
tasks in the pipeline.
126-
</small>
126+
</sub>
127127
</td>
128128
<td>✅</td>
129129
<td>❌</td>
130130
</tr>
131131
<tr>
132132
<td>⭐</td>
133133
<td>
134-
<strong>Nested context extraction</strong>:
135-
<small style="opacity: 0.8;">
134+
<strong>Nested context extraction</strong><br>
135+
<sub>
136136
Automatically manages nested context based on the pipeline definition (e.g.
137137
document > aspects > sub-aspects > concepts).
138-
</small>
138+
</sub>
139139
</td>
140140
<td>✅</td>
141141
<td>⚠️</td>
142142
</tr>
143143
<tr>
144144
<td>⭐</td>
145145
<td>
146-
<strong>Built-in concurrent I/O processing</strong>:
147-
<small style="opacity: 0.8;">
146+
<strong>Built-in concurrent I/O processing</strong><br>
147+
<sub>
148148
Automatically manages concurrent I/O processing to speed up complex
149149
extraction workflows, with a simple switch (<code>use_concurrency=True</code>).
150-
</small>
150+
</sub>
151151
</td>
152152
<td>✅</td>
153153
<td>⚠️</td>
154154
</tr>
155155
<tr>
156156
<td>⭐</td>
157157
<td>
158-
<strong>Fallback and retry logic</strong>:
159-
<small style="opacity: 0.8;">
158+
<strong>Fallback and retry logic</strong><br>
159+
<sub>
160160
Built-in retry logic and easily attachable fallback LLMs.
161-
</small>
161+
</sub>
162162
</td>
163163
<td>✅</td>
164164
<td>✅</td>
165165
</tr>
166166
<tr>
167167
<td>⭐</td>
168168
<td>
169-
<strong>Multiple LLM providers</strong>:
170-
<small style="opacity: 0.8;">
169+
<strong>Multiple LLM providers</strong><br>
170+
<sub>
171171
Compatible with a wide range of commercial and locally hosted LLMs.
172-
</small>
172+
</sub>
173173
</td>
174174
<td>✅</td>
175175
<td>✅</td>
176176
</tr>
177177
<tr>
178178
<td>⭐</td>
179179
<td>
180-
<strong>Usage & costs tracking</strong>:
181-
<small style="opacity: 0.8;">
180+
<strong>Usage & costs tracking</strong><br>
181+
<sub>
182182
Automatically tracks usage (calls, tokens, costs) of all LLM calls.
183-
</small>
183+
</sub>
184184
</td>
185185
<td>✅</td>
186186
<td>✅</td>
187187
</tr>
188188
</tbody>
189189
</table>
190190

191-
✅ - <small>fully supported - no additional setup required</small><br>
192-
⚠️ - <small>partially supported - requires additional setup</small><br>
193-
❌ - <small>not supported - requires custom logic</small>
191+
✅ - <sub>fully supported - no additional setup required</sub><br>
192+
⚠️ - <sub>partially supported - requires additional setup</sub><br>
193+
❌ - <sub>not supported - requires custom logic</sub>
194194

195195
\* See [comparison](https://contextgem.dev/vs_other_frameworks.html) of specific implementation examples using ContextGem and other popular open-source LLM frameworks.
196196

0 commit comments

Comments
 (0)