Examples: query, "exact match", wildcard*, wild?ard, wild*rd
Fuzzy search: cake~ (finds cakes, bake)
Term boost: "red velvet"^4, chocolate^2
Field grouping: tags:(+work -"fun-stuff")
Escaping: Escape characters +-&|!(){}[]^"~*?:\ with \, e.g. \+
Range search: properties.timestamp:[1587729413488 TO *] (inclusive), properties.title:{A TO Z}(excluding A and Z)
Combinations: chocolate AND vanilla, chocolate OR vanilla, (chocolate OR vanilla) NOT "vanilla pudding"
Field search: properties.title:"The Title" AND text
Answered
Hi All, I Have An Issue With The Way Hyper Parameters Are Logged Under Configuration, The Values That Are Stored Seem To Add Unnecessary Escape Characters To The Original Values.. Is It A Known Issue? Is There A Way To Change It? Thanks

Hi all, I have an issue with the way hyper parameters are logged under configuration, the values that are stored seem to add unnecessary escape characters to the original values.. is it a known issue? is there a way to change it? thanks

  
  
Posted 3 years ago
Votes Newest

Answers 30


No worries, I would love for us to come up with a nice solution 🙂

  
  
Posted 3 years ago

AgitatedDove14 , I see, someone must have faced the issue of dumping regular expression strings in tuple before?
I'll have a think and a look too, unfortunately not today

  
  
Posted 3 years ago

DepressedChimpanzee34 any string serialization package I tried will convert r"some\blah" into "some\\blah" (json yaml hocon) otherwise you end up with \b as an escape character. I'm really not sure what to do here. (And reinventing the standard seems unhealthy)

  
  
Posted 3 years ago

DepressedChimpanzee34
so parsing bask is done via a yaml reader:
https://github.com/allegroai/clearml/blob/49fcbd7bbf3236f4175cdff29fa951847b0923cc/clearml/backend_interface/task/args.py#L506
We could add extra test here, checking for \ in the string, that should solve it and will be backwards compatible (I think)
https://github.com/allegroai/clearml/blob/49fcbd7bbf3236f4175cdff29fa951847b0923cc/clearml/backend_interface/task/task.py#L935

  
  
Posted 3 years ago

AgitatedDove14
What I'm looking here is some logic that will allow us not to break backwards compatibility on the one hand, but still will allow you to have something like "first\second" entry. WDYT? any ideas? (I really want to make sure we fix it as soon as possible)Do you mean something like the dictionary structure?
I am not sure about a clean elegant solution yet.. but this patch does some of the job:
str('(' + ', '.join([str(elm) for elm in tuple_value])+')')I am actually curious about the way you cast it back to the original dtype of the individual elements.. because the patch above only solves the display issue as far as I understand

  
  
Posted 3 years ago

BTW:
str('\.') Out[4]: '\\.' str(('\.', )) Out[5]: "('\\\\.',)"This is just python str casting

  
  
Posted 3 years ago

DepressedChimpanzee34 <character> will almost always be converted into \ because otherwise it will not support \t or \n etc.
What I'm looking here is some logic that will allow us not to break backwards compatibility on the one hand, but still will allow you to have something like "first\second" entry.
WDYT? any ideas? (I really want to make sure we fix it as soon as possible)

  
  
Posted 3 years ago

there seem to be an additional logic to:
str_value = str(value)str('.') -> '\.'
however in the configuration I see: '.' for non nested strings

  
  
Posted 3 years ago

ohh actually I think I remember, when you connect a dictionary, the local dtype is used for the casting of the remote matching key (probably more nuanced)

  
  
Posted 3 years ago

I understand that to report any value should be presented as string, how does the "inverse casting" work when I pull some value from the config?

  
  
Posted 3 years ago

It seems to me that the source of the mismatch is the str(tuple())

  
  
Posted 3 years ago

AgitatedDove14 , in my use case, the strings are regular expressions. If the reported regular expression string changes when reported, it messes up my run.
I can't define the "legal regular expression space", off the top of my head. The expected behavior is reported_string == original_string..

  
  
Posted 3 years ago

Hmm DepressedChimpanzee34 my bad it seems the loading is done via YAML loader, but the dumping is straight forward str casting...
https://github.com/allegroai/clearml/blob/6e6271fb91f2aeb2aa7a13c6d07d4e635baaa670/clearml/backend_interface/task/task.py#L934
What would you expect to get (BTW "value\blah" is Not a valid string assignment in python as there is no \b escape character, it should be "value\blah" which translates into the text "value\blah")

  
  
Posted 3 years ago

AgitatedDove14 , agree \ disagree?

  
  
Posted 3 years ago

AgitatedDove14 , when I test using the yaml python package:
I see the following:
import yaml yaml.dump({'\.': ('a', '\.')}) [In]: '\\.: !!python/tuple\n- a\n- \\.\n'YAML treats both strings in tuples and outside the same, however this is not the behavior you get in clearml task.connect

  
  
Posted 3 years ago

so a different behavior between a string and a string in a tuple is by design? I find it confusing, I guess this is the YAML convention?
https://colab.research.google.com/drive/1w5lQGxsblnLGlhJEDH_b0aIiUvjGeLjy?usp=sharing

  
  
Posted 3 years ago

this topic is about the issue with reporting a configuration with a string inside a tuple that has backslash

So the encoding itself is done YAML style, and based on your example \b Has to be encoded to \b because this is string encoding, like \n will become "new line"
Make sense ?

  
  
Posted 3 years ago

(value, "")->(value, "\")

  
  
Posted 3 years ago

well, kind of, I linked the other topic, but it was completely unrelated
this topic is about the issue with reporting a configuration with a string inside a tuple that has backslash

  
  
Posted 3 years ago

I thought this is the issue on the thread you linked, did I miss something ?

  
  
Posted 3 years ago

if preferred I can open a Github issue about this

  
  
Posted 3 years ago

does my explanation makes sense?

  
  
Posted 3 years ago

AgitatedDove14 , by the way, can you take a look at https://clearml.slack.com/archives/CTK20V944/p1625558368001600
maybe you'll have other ideas? at the moment it seems like a dead end

  
  
Posted 3 years ago

I am actually not sure specifically about \b myself, but even when replacing with . I am getting \. double backslash instead of the single backslash ( for the tuple case ). which in the case of a regexp expression changes the meaning of the expression. the expected behavior would be registering it as single backslash

  
  
Posted 3 years ago

Sorry found the code on the Task, duh 🙂
# get_ipython().magic('pip install clearml') import clearml from clearml import Task task = Task.init(project_name='examples', task_name='test param', reuse_last_task_id=False) param = { 'tuple_double_quotes_r': (r"value\blah", 1), 'tuple_double_quotes': ("value\blah", 1), 'tuple_single_quotes': ('value\blah', 1), "double_quotes_r": r"value\blah", 'double_quotes': "value\blah", 'single_quotes': 'value\blah' } task.connect(param) print('done') r"value\blah"what are you expecting to get here ? '\b' is not a valid escaping (I think)

  
  
Posted 3 years ago

Hi DepressedChimpanzee34
How do I reproduce the issue ?
What are we expecting to get there ?
Is that a Colab issue or hyper-parameter encoding issue ?

  
  
Posted 3 years ago

It would be very very useful for my use case, and I believe a relatively popular use case in general for example when using regular expression configurations

  
  
Posted 3 years ago

It should preserve the order as the order of the update back (i.e. when executed by the agent) is the same as the order of the keys (obviously py3.7+ becuase it creates dict not Ordered Dicts)

  
  
Posted 3 years ago

mostly the second case

  
  
Posted 3 years ago
928 Views
30 Answers
3 years ago
one year ago
Tags