cutright / dvh-analytics-bokeh Goto Github PK

see title

surface area calculation for ROIs has been rewritten for calculation at time of DICOM import here: 0bd6268

• Need to add ability to recalculate surface areas of ROIs already in database
• Need to validate calculations against a TPS or maybe CERR

see title

Hi Cutright,

Thanks for quick response. I almost there, but still have issues to run as shown below. By the way, where the import directories (review, inbox,imported) are created by default?

Thanks,

tom

C:\ProgramData\Anaconda2\lib\site-packages\fuzzywuzzy-0.15.1-py2.7.egg\fuzzywuzzy\fuzz.py:35: UserWarning: Using slow pure-python SequenceMatcher. Install python-Levenshtein to remove this warning
Loading create_tables.sql
Executing create_tables.sql
Tables created
True
2017-10-18 13:40:58,976 Starting Bokeh server version 0.12.7 (running on Tornado 4.4.2)
2017-10-18 13:40:58,984 Bokeh app running at: http://localhost:5006/dvh
2017-10-18 13:40:58,984 Starting Bokeh server with process id: 9532
2017-10-18 13:41:04,216 Error running application handler <bokeh.application.handlers.directory.DirectoryHandler object at 0x0000000008FB8710>: list index out of range
File "main.py", line 496, in init:
self.select_value = Select(value=self.values[0], options=self.values, width=420) Traceback (most recent call last):
File "C:\ProgramData\Anaconda2\lib\site-packages\bokeh\application\handlers\code_runner.py", line 125, in run
exec(self._code, module.dict)
File "C:\AitangResearch\DVHAnalytics\dvh\main.py", line 3487, in
button_add_selector_row()
File "C:\AitangResearch\DVHAnalytics\dvh\main.py", line 220, in button_add_selector_row
query_row.append(SelectorRow())
File "C:\AitangResearch\DVHAnalytics\dvh\main.py", line 496, in init
self.select_value = Select(value=self.values[0], options=self.values, width=420)
IndexError: list index out of range

2017-10-18 13:41:04,226 200 GET /dvh (::1) 4141.00ms
2017-10-18 13:41:04,430 304 GET /dvh/static/images/logo.jpg (::1) 162.00ms
2017-10-18 13:41:04,555 WebSocket connection opened
2017-10-18 13:41:04,555 ServerConnection created
2017-10-18 13:41:04,585 404 GET /favicon.ico (::1) 0.00ms

An empty return after a query results in an error on the server terminal, but the user at the browser is unaware

see title

error handling message Message 'PATCH-DOC' (revision 1): AttributeError("'FileDataset' object has no attribute 'BeamSequence'",)

Time Series DVH EP choices are updated properly, but Regression DVH EP choices are not.

It seems data may not be removed from source_endpoint_calcs when deleting or changing DVH EP type? Regression plot can view old data not shown in DVH EP table or Time-Series.

The MLC Analyzer module was designed with Pinnacle DICOM files. I recently tested with a RayStation plan and it crashed. Modifications need to be made for other TPSs.

Hi,

I found the institutional ROI and physician ROI are all uncategorized after data importing, so I go to the ROI name manager to remap ROIs. But there is a error:

error handling message Message u'EVENT' (revision 1) content: u'{"event_name":"button_click","event_values":{"model_id":"2221"}}': KeyError('',)

I don't know if it is a bug or what, could you please give me some hints? Thanks!

Best,

Zeyu

Would be nice to have potential issues with query posted to help explain why certain modules may be failing. For example:

• Non-numerical data detected for numerical SQL columns
• Missing data (e.g., common ones could be simulation date, age, physician, etc.)
• Missing PTV in a plan
• Missing PTV Distance and PTV Overlap calculations

Developing a UI for options.py will make it easier for users to customize, especially Docker users.

see title

There is a bug in Bokeh such that data tables may not render if tabs are used, and the data table source changes. An issue was filed here: bokeh/bokeh#6925

Issue still persists in Bokeh 0.12.10

Add functionality for rolling average by # days or # patients

I noticed trying to import some data I have on hand that the importer assumes that the Beam Dose Specification Point is set in the RTPLAN file. The exact lines of code that handle that are here. As far as I understand, this isn't required by the DICOM specification, and as such there is no guarantee this is set (note that I could easily be wrong here as I'm not a DICOM expert). At the moment any RTPLAN files without this attribute will crash the importer. I'm not sure what exactly the point is used for elsewhere in the program for analysis, or what should be stored in the database if the point isn't present.

This is been a known bug for a long time, it's more of a nuisance than a bug. Does anyone know how to turn off or cancel the browser scrolling in javascript? I don't believe Bokeh provides explicit methods for this.

There's no need to set this option in options.py... would be much easier to just have a checkbox

see title

Create a default ROI map based on TG-263
https://www.aapm.org/pubs/reports/RPT_263.pdf

The count reported in the Query button of the Query tab does not appear to account for group assignments.

It is unclear which instance is being deleted. Either delete all for a given physician, or indicate which one will be deleted.

Adding a new endpoint will update to the correlation data, but editing or deleting don't appear to properly update in the correlation data used for the Correlation and Regression tabs

Error:

AttributeErrorTraceback (most recent call last)
<ipython-input-42-a660279469d7> in <module>()

from dvh import mlc_analyzer as mlca
----> 1 pcs = mlca.Plan(rtplan1_dcm)

C:\AppData\Local\conda\conda\envs\mppy27\lib\site-packages\dvh\mlc_analyzer.pyc in __init__(self, rt_plan)
     16 class Plan:
     17     def __init__(self, rt_plan):
---> 18         self.fx_group = [FxGroup(fx_grp_seq, rt_plan.BeamSequence) for fx_grp_seq in rt_plan.FractionGroupSequence]

AttributeError: 'str' object has no attribute 'FractionGroupSequence'

Although rtplan.ds.FractionGroupSequence does exist.
Please, how do i get around this.
Thank you.

Would be nice to download the query, but anonymize the data

If import calculations change in the code, would be nice to reprocess specific calculations from DICOM without having to completely reimport a plan.

Hi Cutright,
Thanks a lot for developing this great tools. When I setup and try to run the main.py, the following errors occurred. Could you help me to resolve this? Thanks.

tom

File "C:\AitangResearch\DVHAnalytics\dvh\main.py", line 50, in
temp_dvh_info = Temp_DICOM_FileSet()
File "C:\AitangResearch\DVHAnalytics\dvh\utilities.py", line 31, in init
with open(abs_file_path, 'r') as document:

IOError: [Errno 2] No such file or directory: 'C:\AitangResearch\DVHAnalytics\dvh\preferences/import_settings.txt'

Calc all PTV to OAR distances, rather than just categorized ROIs. If ROI point count greater than some threshold, downsample and calc.

This likely only affects external and skin ROIs, or similar, which added error in absolute min distance would be relatively small.

Fix coming in 0.4.3

Would be nice to keep as is, if switching between variables from the same table, but should reset to whole data set if switching to new table?

Look into auto update of trending (requires sensing mouse button status)

The table for 'Query by Numerical Data' cannot handle dates, query fails. Need to adjust update_range_source() to handle dates:
https://github.com/cutright/DVH-Analytics/blob/6bbb4d360ae10814caa6eb11295b10f105fad619/dvh/main.py#L385-L410

Need to exclude data from the DVH source that was added for max, min, mean, etc. when trending in the Range-Variable Trending

We intend to add automation of composite DVH calculation based on multiple DICOM RT Dose files

Because not all treatment planning systems can export a composite plan (e.g., putting all rxs in a single rt plan file), functionality to import all rt plan files for a given study instance UID was added (see 3e3db72).

Currently, DVH Analytics will still only import the latest dose and struct files for a given study instance UID, so the user may need to use other software to generate a single rt dose file of the composite dose. We plan to do this for you in a future update.

In summary, DVH Analytics assumes:

• All rt plan files in the inbox are pertinent (unless IMPORT_LATEST_PLAN_ONLY is set to True in options.py)
• Only the latest rt plan and rt struct files will be used, and are assumed to reflect a composite
  • All DICOM files for a given study instance UID will still be catalogued, even if their data is not added to the Plans, Beams, Rxs, or DVHs tables

Hi

First of all, thank you for this extremely useful tool.
I am tried to feed in Pinnacle and Eclipse treatment plans into the system and strangely experienced a shift in beam MUs at the eclipse plans. I have checked the dcm files and the beam numbers are correct and referencing the correct amount of MUs. Have you come cross this issue? Is there a compatibility issue with eclipse versions?
We don`t have problems with the Pinnacle plans, those are coming through fine.

Thank you

i am very insterseted in it for my reserach,but there are many componennts to install after configuring python 2.7 ,can you describe the simplified installation procedure？thanks

In __main__.py, Temp_DICOM_FileSet is still referenced from dvh/utilities.py instead of dvh/bokeh_components/utilities.py

__main__.py is the entry point for pip installed versions of DVH Analytics, therefore 0.4.6 is not runnable via 'dvh' calls from a pip install or in docker.