File size: 5,132 Bytes
0914710
 
a1d837d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
bb5a051
 
 
 
 
 
 
a1d837d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0914710
bb5a051
0914710
 
 
bb5a051
0914710
 
 
bb5a051
0914710
 
 
 
bb5a051
0914710
 
 
 
 
 
 
bb5a051
0914710
 
 
 
 
 
bb5a051
0914710
 
 
bb5a051
0914710
bb5a051
0914710
 
 
 
 
 
 
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
# Changelog

## Version 1.5.1

chore: update samgis-lisa == 1.0.2, bump samgis-lisa-on-zero to version 1.5.1 alessandro trinca tornidor Moments ago
feat: fix fastapi app title alessandro trinca tornidor 13/08/24, 23:58
refactor: remove unused files, update samgis-lisa to version 1.0.1 also in pyproject.toml alessandro trinca tornidor 07/08/24, 19:13
feat: use new optional argument inference_decorator=spaces.GPU for lisa.lisa_predict() alessandro trinca tornidor 07/08/24, 19:12
chore: update samgis-lisa to version 1.0.1 alessandro trinca tornidor 07/08/24, 19:11

## Version 1.5.0

- bump to version 1.5.0, remove unused direct dependencies, add note about huggingface space dependencies
- use new optional argument inference_decorator=spaces.GPU for lisa.lisa_predict()
- update submodule sam-quantized
- update samgis-lisa to version 1.0.1
- update samgis-core to version 3.0.12, add samgis-lisa version 1.0.0, samgis-web version 1.0.13
- update gradio to version 4.40.0
- update missing lfs images
- refactor moving code to separate python package samgis-lisa
- rename io package to 'io_package' to avoid errors on test execution. Add pytest dependency
- avoid 404 for output.css on loading of custom html pages, handle lfs images the right way
- remove loguru and change logging method (use the @session_logger.set_uuid_logging decorator on functions with logs)
- format gradio app into row and columns
- add gradio markdown support
- fix infer_lisa() because of the body dumped to json two times
- remove unuseful app_helpers.get_inference_model_by_args(args, inference_decorator=SPACES_GPU), fix logs, initialize gpu within infer_lisa_gradio()
- handle custom relative url with env variables also on the frontend
- add missing onnx model files for mobile-sam used by standard samgis page
- add custom "/static" mount folder used by custom pages
- change gradio url to /
- remove unused routes /post_test_dictlist, /post_test_string, /infer_samgis
- handle custom urls with env variables: CUSTOM_INDEX_URL, CUSTOM_SAMGIS_URL, CUSTOM_LISA_URL, CUSTOM_GRADIO_URL
- rename wrong references to samgis_lisa_on_cuda in urls and page descriptions
- add missing statements for gradio/fastapi app
- add missing images

## Version 1.4.0

feat: first attempt to support samgis-lisa-on-cuda on ZeroGPU huggingface space
- add source to prefix string on WRITE_TMP_ON_DISK env
- fastapi_wrapper.py: handle multiple exceptions the right way
- handle also IsADirectoryError as exception in case of WRITE_TMP_ON_DISK env
- improve logs during docker build
- permit download of tmp images if WRITE_TMP_ON_DISK env exists

## Version 1.3.3

- refactor UI
- switch to base docker image pytorch/pytorch:2.2.2-cuda12.1-cudnn8-runtime
- now LISA+SamGIS demo is the default demo
- lock map on inference to take advantage of image embedding re-use

## Version 1.3.2

- update lisa-on-cuda dependency to version 1.0.4

## Version 1.3.1

- copy resources folder from lisa-on-cuda repository to /var/task/resources, bump to version 1.3.1
- fix wrong venv, create_folders_and_variables_if_not_exists.py paths in entrypoint.sh
- fix entrypoint.sh exec permission
- add scripts folder, update README.md
- use entrypoint.sh as CMD Dockerfile, add FOLDERS_MAP env file to dynamic folder creation

## Version 1.3.0

- rename samgis module to samgis_lisa_on_cuda, bump to version 1.3.0, update samgis == 1.0.7, lisa_on_cuda == 1.0.2 dependencies
- refactor base docker image
- dockerfile-lisa-base: remove workaround for old versions of docker
- update direct URL space
- fix /health

## Version 1.2.3

- sync with samgis-be (pyproject.toml, PageLisaMap.vue), update HF space card title
- pyproject.toml: update samgis-core, lisa-on-cuda git urls
- Dockerfile, poetry install: remove verbose option
- Dockerfile: add ARG DEPENDENCY_GROUP=fastapi
- handle SAM machine learning models folder

## Version 1.2.2

- update dependencies version

## Version 1.2.1

- update dependencies version

## Version 1.2.0

- code refactor to separate core functionality (instance segmentation) from other code
- updated test coverage

## Version 1.1.0

- Added this changelog
- specific backend branch code uses terrain providers like nextzen and MapBox Terrain-RGB v1
- update test coverage
- update python dependencies versions
- update node dependencies versions

## Version 1.0.2

- HuggingFace frontend demo: update the navbar url to SamGIS docs
- add two openapi docs builder scripts
- update samgis version within docs
- remove unused package.json

## Version 1.0.1

- fixed zlib1g security vulnerability in python:3.11-slim-bookworm docker image, see https://security-tracker.debian.org/tracker/CVE-2023-45853

## Version 1.0.0

First release:

- user onboarding tour with driver.js
- code decoupling between the AWS lambda wrapper and the backend...
- ...now also deployed on this HuggingSpace space demo (here some explanation about adding a SPA vuejs frontend)
- request input validation using Pydantic
- support for array prompts (both rectangle and point types)
- tiles download/merge/crop steps uses contextily by geopandas
- CSS frontend style with Tailwind