All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog and this project adheres to Semantic Versioning.
v0.15.2 - 2021-02-02
- Wrong calc in crophook (72d47df by williamfzc).
v0.15.1 - 2020-11-07
- Change default values of keras train (c44b6fd by williamfzc).
- Verify dataset before train (cb78e27 by williamfzc).
v0.15.0 - 2020-09-07
- Remove deprecated api (967a08c by williamfzc).
- Ignore_error in calc (6ed342b by williamfzc).
- Path check first (504667b by williamfzc).
- Calcconfig (8043f57 by williamfzc).
- Api
contain(b6e5d05 by williamfzc). - Remove
pre_loadand useload_framesinstead (3e1f673 by williamfzc). - Enable video preload by default (dd2adec by williamfzc).
- Extraconfig (7f2517c by williamfzc).
- Videoconfig (98cb9e6 by williamfzc).
- Custom
compare_frame_listmethod (d215134 by williamfzc). runmethod accepts a path or a preload dict (e64b983 by williamfzc).- Run with config file (829659b by williamfzc).
v0.14.2 - 2020-08-06
- Support target_size in
draw(28cd4dc by williamfzc). - Reversed size (802bd98 by williamfzc).
- Hook should always be applied in all the frames (3184a0e by williamfzc).
- Size and count should be updated after hook (2d28a63 by williamfzc).
v0.14.1 - 2020-07-18
- Support preload hook for video object (c7af279 by williamfzc).
v0.14.0 - 2020-06-12
- Do not wrap too much (875ae42 by williamfzc).
- Report path ends with .html (e5875a9 by williamfzc).
- Report path can be a directory (ae93a2b by williamfzc).
- Set parallel to 3 (d035fda by williamfzc).
- Stable api (51b239a by williamfzc).
v0.13.1 - 2020-05-25
- Very effective boost mode for all classifiers (2d1cf1c by williamfzc).
v0.12.1 - 2020-04-27
- Add warning in report (ad66a90 by williamfzc).
- Save trained model to random path if existed (1e09e6d by williamfzc).
v0.12.0 - 2020-04-02
- Error when step != 1 (4de619a by williamfzc).
v0.11.2 - 2020-03-22
- Unstable and unspecific (b3eea51 by williamfzc).
- Type error (6fc668c by williamfzc).
- Pathlib issue in videoobject (b11041d by williamfzc).
- New
helpertab for report.html (1406f68 by williamfzc). - Add dump() for classifierresult (212bb91 by iceyhuang).
v0.11.1 - 2020-03-10
- Time_cost_between() for quick calc (6fd7460 by williamfzc).
v0.11.0 - 2020-03-02
- Error if video has an unstable end (b1e222d by williamfzc).
- Final stage (1e6adc1 by williamfzc).
- Get_stage_range (46d2e46 by williamfzc).
- Timestamp (c6dc3ad by williamfzc).
- About how pointer moving between frames (3329eb9 by williamfzc).
- Pointer offset (b2be6f2 by williamfzc).
- Stage range error (4511213 by williamfzc).
- Meaning of timestamp (22aef44 by williamfzc).
- Add extra frame in unstable range (59c0815 by williamfzc).
- Default threshold 0.95 -> 0.98 (2259727 by williamfzc).
v0.10.4 - 2020-02-26
- Wrong attr name (078d12b by williamfzc).
- Api for unstable range (38ef721 by williamfzc).
- More api (4fdda5c by williamfzc).
v0.10.3 - 2020-02-19
- Add tensorflow (8c0bc44 by williamfzc).
- Avoid conflict (178c677 by williamfzc).
- Lock opencv-contrib-python version (7bfc2be by williamfzc).
- Remove surf (9f60be9 by williamfzc).
- Keras train function in command cli (022eaa4 by williamfzc).
- Add some warning msg when saving frames to a existed dir (d678d7b by williamfzc).
- Bump findit's version to 5.8 (29d068e by williamfzc).
v0.10.2 - 2020-01-31
- Raise exception when dir existed (6408e26 by williamfzc).
v0.10.1 - 2020-01-19
- Use notimplementederror in abstract methods (39e0455 by williamfzc).
- Length check (038a829 by williamfzc).
- Meaningful image names (d70ac78 by williamfzc).
v0.10.0 - 2020-01-12
- Wrong way to use model (1e63c49 by williamfzc).
- Use sparse_categorical_crossentropy (cc4f737 by williamfzc).
- Turn_binary (bcd585c by williamfzc).
- Keras setup (f02286d by williamfzc).
- Model design (498e949 by williamfzc).
- Add basemodelclassifier for extending (1f996fd by williamfzc).
- Basic keras classifier (e9059cf by williamfzc).
v0.9.4 - 2020-01-02
- Handle hooks at the beginning (6c4c612 by williamfzc).
- New api for important frames (72b2d03 by williamfzc).
v0.9.3 - 2019-12-26
- Remove
overwrite(f9abb41 by williamfzc).
- Better file name for framesavehook (4740bbf by williamfzc).
- Better hook design (00193dc by williamfzc).
v0.9.2 - 2019-12-17
- Resize the video length after load_frames (1817cb7 by williamfzc).
- For actions/virtual-environments (9ef6a3d by williamfzc).
- Better support for findit (ac027b1 by williamfzc).
- 3.7.5 > 3.7 in setup.py (f75ae67 by williamfzc).
- To_dict return ordereddict (e736642 by williamfzc).
v0.9.1 - 2019-12-05
- Error case (b3f2541 by williamfzc).
- Contain_image allow kwargs only (for extra args in findit) (1c9f989 by williamfzc).
- Move match_template to toolbox.py (bb06077 by williamfzc).
- Support contain_image in singleclassifierresult (d9aee84 by williamfzc).
- Extra args (optional) for contain_image (ae04813 by williamfzc).
- Support contain_image in videoframe (16b70eb by williamfzc).
v0.9.0 - 2019-11-22
- Error assertion (7d2b1b9 by williamfzc).
- Remove unused cases (a0373f9 by williamfzc).
- Duration is hard to understand (6a29656 by williamfzc).
- Ffmpeg need
sudo(a80ef82 by williamfzc). - Install ffmpeg firstly (2cae8c4 by williamfzc).
- Put all the flags to constants.py (6dcac6a by williamfzc).
- New
drawfunction (c6fcf32 by williamfzc). - Render with classifier result only (864c178 by williamfzc).
- Video info in report (5f37bb5 by williamfzc).
- New design for easily review :) (1dc857f by williamfzc).
- Support portable ffmpeg (2c0a5e7 by williamfzc).
- Ffmpeg converter (df13699 by williamfzc).
v0.8.4 - 2019-11-18
- Flexible blocks (fbd0ab9 by williamfzc).
v0.8.3 - 2019-11-05
- Pyecharts can not handle float list directly (29f3d32 by williamfzc).
- Miss thumbnail in mini.py (bf250ac by williamfzc).
- Video_jump does not lead to correct point (e6e78a4 by williamfzc).
v0.8.2 - 2019-10-26
- Sync example (5cff6ba by williamfzc).
- Sync test cases (dce9edf by williamfzc).
- Lock python image version to 3.7 (d54d509 by williamfzc).
- Opencv-python version (43a14a1 by williamfzc).
- Lock opencv-python version to 4.1.0.25 (0a8e8e1 by williamfzc).
- Output of classifier has became an classifierresult object, not a list (6cec3e8 by williamfzc).
v0.8.1 - 2019-10-14
- Skimage version 0.14.2 to 0.16.0 (c85d62d by williamfzc).
- Dead loop occasionally (d57a29c by williamfzc).
- Rename emptyframedetecthook to interestpointhook (be24630 by williamfzc).
- Type warning from
pyright(aeb81a5 by williamfzc).
- Add pre_load option in init (9f34692 by williamfzc).
- Support emptyframedetecthook (ba8ef2a by williamfzc).
- Add version code at the footer of report (175d855 by williamfzc).
v0.8.0 - 2019-09-26
- Error import (665f2ca by williamfzc).
- Avoid import loop (78dc9b3 by williamfzc).
- Bug found by unittest (43032f2 by williamfzc).
- Error because of numpy.ndarray -> json (e9d4dc4 by williamfzc).
- Cutter (9fb3377 by williamfzc).
- Support videoobject in api.py (c4a0182 by williamfzc).
- Support videoobject in cut and classify (25e26b4 by williamfzc).
- Add videoframe / videooperator / videoobject (41c9d68 by williamfzc).
v0.7.5 - 2019-09-24
- Always cut before classify (3d83c5d by williamfzc).
- Classify with trained model failed (8f7dc18 by williamfzc).
- Miss
trainin all (8c791ad by williamfzc).
v0.7.4 - 2019-09-24
- Model name conflict (9d55e92 by williamfzc).
- Add test cases for
trainin api (c009dcc by williamfzc).
- Less code, better structure (5605a1e by williamfzc).
- Split cli.py into 2 files (815f899 by williamfzc).
- Split cli.py into cli and api (ded8a73 by williamfzc).
- Support
train(af05c72 by williamfzc). - Support classifying with a trained model (98411ef by williamfzc).
- No need (495107e by williamfzc).
v0.7.3 - 2019-09-17
- Support chinese report (7a9656f by williamfzc).
- Support
saveandloadin reporter (b19ff7d by williamfzc).
v0.7.2 - 2019-09-10
v0.7.1 - 2019-09-10
- Replace
cv2.imreadwithtoolbox.imread(f360212 by williamfzc). - Do not run cli in shell (697129a by williamfzc).
- (try to) using py.test with coverage doesn't include imports (bc8ec3d by williamfzc).
- Video path (e601912 by williamfzc).
- Github actions pipeline (c6ae64a by williamfzc).
- Case structure improvement (2c51277 by williamfzc).
- Support ignorehook! (3f81c77 by williamfzc).
v0.7.0 - 2019-09-08
- In the binary case, return type is different (wtf ...) (a8de2bd by williamfzc).
- Default thumbnail maybe different (no args and kwargs) (127193d by williamfzc).
- Support crophook! (68a4b2f by williamfzc).
- Add duration data in thumbnail (8d70fa2 by williamfzc).
- Support score_threshold (b2d5ba7 by williamfzc).
- Add scores of each stages in log (cc899a7 by williamfzc).
v0.6.2 - 2019-09-04
v0.6.1 - 2019-09-01
- Default compress rate (feb4351 by williamfzc).
- Default compress rate is none (dc3c850 by williamfzc).
- Better tutorial (8cd3e9d by williamfzc).
- Threshold in kwargs (b011073 by williamfzc).
- Test for cli (2d87c2d by williamfzc).
- Support base classifier api (0991505 by williamfzc).
- Finish cut api (9592475 by williamfzc).
- Finish one-step api (7bacd03 by williamfzc).
- Add cli test in workflow (24745ed by williamfzc).
- Base cli for terminal! (9d2bb54 by williamfzc).
- Limit in stable ranges and unstable ranges (fea1d85 by williamfzc).
v0.6.0 - 2019-08-30
- Extra kwargs will not be used in diff (689965d by williamfzc).
- When err == 0, psnr will be 'inf' (d0e11f7 by williamfzc).
- Background color in report (e1065d1 by williamfzc).
- Report improvement (b6d8dec by williamfzc).
- Disable is_step (0ccb3c5 by williamfzc).
- Time cost of stage changing (fa47ba0 by williamfzc).
- Add split line in thumbnail (7218dfd by williamfzc).
- More info in return of diff (3eed5bd by williamfzc).
- Built-in thumbnail (d3aaeeb by williamfzc).
v0.5.3 - 2019-08-25
- Github actions badge (d96dbf2 by williamfzc).
- Set ssim + psnr as default (b51dc00 by williamfzc).
- Add psnr (42309c2 by williamfzc).
- Add sim chart (78e8ec6 by williamfzc).
- Add mse (bd8286e by williamfzc).
v0.5.2 - 2019-08-22
- Better sampling algorithm (b512fa8 by williamfzc).
v0.5.1 - 2019-08-18
- Move all the frame-level operations to hook (4fda59a by williamfzc).
- Support label 'overwrite' in hook (694319a by williamfzc).
v0.5.0 - 2019-08-16
- Miss last stage (75a7ab1 by williamfzc).
- Timestamp error (f08a467 by williamfzc).
- Remove data assertion before classify (e7a4497 by williamfzc).
- Support pruning redundant stages (24cddc9 by williamfzc).
- Support findit hook (d1d8340 by williamfzc).
- Add option 'auto_merge' in diff function (af7123d by williamfzc).
v0.4.3 - 2019-08-14
- Support diff in videocutresult (8a014ae by williamfzc).
- Support videoframe (050efd1 by williamfzc).
- Support load and dump (913d78f by williamfzc).
v0.4.2 - 2019-08-11
- Support invalid frame detect hook (3fa8026 by williamfzc).
- Support framesavehook (0253304 by williamfzc).
- Support hook (c3521a2 by williamfzc).
v0.4.1 - 2019-08-09
- Use videoobject, not video path (721503f by williamfzc).
v0.4.0 - 2019-08-06
- Error when unstable range count equal or less than 1 (bb002bf by williamfzc).
- Index out of range (635aa18 by williamfzc).
- Range_threshold should be optional (b2c6796 by williamfzc).
- Error when video starts (ends) with changing (dc1b208 by williamfzc).
- Error when using compression (dc3926a by williamfzc).
- Too much code in cutter.py (cd44b09 by williamfzc).
- Block cutter (3ecb828 by williamfzc).
- More precise comparison in cutter (8b71435 by williamfzc).
- Support comparison between different results of cutter (87a378b by williamfzc).
- Support offset (0a693bc by williamfzc).
- Support range_threshold (8275535 by williamfzc).
- Change self.ssim to a list of ssim value (271a451 by williamfzc).
- Need not get timestamp from video again (1452525 by williamfzc).
- Timestamp in videocutrange (feeaa67 by williamfzc).
v0.3.3 - 2019-08-01
- Support specific size in compression (f95a4df by williamfzc).
v0.3.2 - 2019-07-28
- Custom content in report (7c1ec71 by williamfzc).
v0.3.1 - 2019-07-26
- Always drop last range before (d50d91e by williamfzc).
- When length == 1, start > end (d62d402 by williamfzc).
- Put html content in script (4327b17 by williamfzc).
- Include .html file in setup.py (9a45366 by williamfzc).
- Better report with bootstrap4 (b10104d by williamfzc).
v0.3.0 - 2019-07-25
- Remove intersection between stable and unstable (91ad621 by williamfzc).
- Relationship between frame id and actual frame (74da7c1 by williamfzc).
- Support inserting thumbnail into report (fe9a0e3 by williamfzc).
- Support saving thumbnail to file (7df155c by williamfzc).
- Thumbnail for stage (6761ff5 by williamfzc).
- Add example (97ceeb9 by williamfzc).
- Support loading range as data directly (7ff66f5 by williamfzc).
v0.2.1 - 2019-07-22
- Support threshold in getting stable (and unstable) range (70afa81 by williamfzc).
v0.2.0 - 2019-07-22
- Use abs path of picture (32b166a by williamfzc).
- Remove un-runnable code (12eb5f6 by williamfzc).
- Lost last frame (09f1936 by williamfzc).
- Calculation of time cost (db54b84 by williamfzc).
- Rename 'none' to 'raw' (4586826 by williamfzc).
- Duplicate code in classifiers (3f96e73 by williamfzc).
- Split different classifiers (595d051 by williamfzc).
- Better reporter (985377d by williamfzc).
- Support 'step' in classify (d828b40 by williamfzc).
- Support findit (eec0f25 by williamfzc).
- Support limit range (3b37623 by williamfzc).
- Support feature extraction (cacecff by williamfzc).
v0.1.0 - 2019-07-17
- Calculation based on all pictures, not one of them (4fc99c4 by williamfzc).
- Syntax issue (f7fbae2 by williamfzc).
- Import and export for model (51f89e1 by williamfzc).
- Better chart (6bbb46a by williamfzc).
- Svm classifier (not precise enough now) (39a214c by williamfzc).
- Chart improvement (31dedc0 by williamfzc).
- Length filter (6b00efb by williamfzc).
- Reporter (6e56bc6 by williamfzc).
- Classify (bcbdabb by williamfzc).
- Cutter (a238a0d by williamfzc).
- Cutter (base) (d2b610a by williamfzc).