| Commit message (Collapse) | Author | Age | Files | Lines |
| | |
|
| |\ |
|
| | | |
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| | |
- Cert_ids are now extracted
- cert_id from frontpage is always preferred
- If no frontpage cert_id is found, the most ocurring keyword is preferred
- Some tests were revoked as heuristics should be computed after pdf processing
|
| | | |
|
| | | |
|
| | | |
|
| |/ |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
| |
- Stored as a part of InternalState complex variable in json
|
| |
|
|
|
| |
- Previously as a part of dataset where certs changed to dictionary
- For rendering at webserver, list is more convenient
|
| | |
|
| |
|
|
|
|
| |
- The variable checks which actions were already performed
- Dataset now checks if someone attempts to e.g. convert pdf->txt prior to downloading
- Also, operations on certificates are performed only on those certs for which the preceeding operation was successful
|
| |
|
|
|
|
|
| |
- Status active or archived held as a string variable
- Instead of implicit computation from not-valid-before/after
- The status is computed from the filename of the csv/html source
- The status is serialized into json
|
| |
|
|
|
|
|
|
| |
- Tests for txt file processing still missing
- The internals of the txt processing methods are still authored by petrs
- They are however fully parallel and called in an OOP fashion
- Overall, all results are gathered in a single json file
- All results are implicitly paired
|
| |
|
|
|
|
| |
- Every certificate now implements a static method for its processing
- These are called in parallel
- There's some code duplication at the moment, stuff can be simplified.
|
| |
|
|
|
|
|
|
| |
- Every complex serialiazable object now inherits from `ComplexSerializableType`
- All such objects now implement `to_dict()`, `from_dict()` methods
- Some objects for which standalone json file makes sense implement `to_json()`, `from_json()`
- Rewrite test and demos
- Datasets `root_dir` is now not part of serialization, but only internal representation
|
| |
|
|
|
|
|
| |
It seems that pdftotext is not fully reproducible and gives platform
specific results. So we dont check for full file equality, but rather
for the existence of converted file. We further check that the created
file size roughly matches the one freshly produced by Travis pdftotext.
|
| | |
|
| | |
|
| |
|
|
| |
The test assumes that the files download correctly.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* FIPS Certificate constructor
* Parsing of single certificate done
* Downloading files
* Create dataset of objects
* demo + dumping + keyword extraction
* wait for approval + everything changed
* non trivial amount of changes to make this work
* modified tests to work with new fips html files
* merge comments resolved (first commit)
* merge comments resolved (second commit)
* realizing a better idea is to leave it alone
* commit #3
Co-authored-by: Stanislav Boboň <xbobon@fi.muni.cz>
|
| | |
|
| | |
|
| | |
|
| |
|