Refactoring to separate file_context elements. Updates to ranks.py

This commit is contained in:
filmfreedom-org 2021-06-08 20:49:32 -05:00
parent 6ff5b6b4f5
commit 43a7adb379
18 changed files with 1392 additions and 1211 deletions

77
abx/enum.py Normal file
View File

@ -0,0 +1,77 @@
# enum.py
"""
A custom enumeration type that supports ordering and Blender enum UI requirements.
"""
class Enum(dict):
"""
List of options defined in a two-way dictionary.
"""
def __init__(self, *options):
"""
Args:
*options (list): a list of strings to be used as enumerated values.
"""
for i, option in enumerate(options):
if isinstance(option, list) or isinstance(option, tuple):
name = option[0]
self[i] = tuple(option)
else:
name = str(option)
self[i] = (option, option, option)
self[name] = i
if name not in ('name', 'number', 'options'):
setattr(self, name, i)
@property
def options(self):
"""
Gives the options in a Blender-friendly format.
Returns:
A list of triples containing the three required fields for
Blender's bpy.props.EnumProperty.
If the Enum was initialized with strings, the options will
contain the same string three times. If initialized with
tuples of strings, they will be used unaltered.
"""
options = []
number_keys = sorted([k for k in self.keys() if type(k) is int])
return [self[i] for i in number_keys]
def name(self, n):
"""
Return the name (str) value of enum, regardless of which is provided.
Args:
n (str, int): An enum value (either number or string).
Returns:
Returns a string if n is recognized. Returns None if not.
"""
if type(n) is int:
return self[n][0]
elif type(n) is str:
return n
else:
return None
def number(self, n):
"""
Return the number (int) value of enum, regardless of which is provided.
Args:
n (str, int): An enum value (either number or string).
Returns:
Returns a number if n is recognized. Returns None if not.
"""
if type(n) is str:
return self[n]
elif type(n) is int:
return n
else:
return None

File diff suppressed because it is too large Load Diff

337
abx/name_context.py Normal file
View File

@ -0,0 +1,337 @@
# name_context.py
"""
NameContext defines the context of a particular named asset.
Examples include scenes in a Blender file, but could be applied to other things.
Also a base class for FileContext which applies to the whole file.
NameContext handles the unified needs of these types, and in particular,
hosts the methods used to generate names.
"""
import os, re
import yaml
from .name_schema import FieldSchema
class NameContext(object):
"""
Single naming context within the file (e.g. a Blender scene).
NameContext defines the characteristics of any particular project
unit (taxon) within the project. So, for example, it may represent
the context of an "Episode" or a "Sequence" or a "Shot".
Used in Blender, it will typically be used in two ways: one to represent
the entire file (as the base class for FileContext) and one to represent
a particular Blender scene within the file, which may represent a single
shot, multiple shots with the same camera, or perhaps just an element
of a compositing shot created from multiple Blender scenes.
Examples of all three uses occur in "Lunatics!" episode 1: the
"Press Conference" uses multicam workflow, with scenes for each camera
with multiple shots selected from the timeline, using the VSE; most scenes
are done 'single camera' on a shot-per-scene basis; but some shots use a
'Freestyle camera clipping' technique which puts the line render in a
separate (but linked) scene, while the final shot in the episode combines
three different Blender scenes in a 2D composite, to effect a smooth
transition to the titles.
Attributes:
container (NameContext):
The project unit that contains this one. One step
up the tree, but not necessarily the next step up
in rank (because there can be skipped ranks).
schemas (list(FieldSchema)):
The schema list as seen by this unit, taking into
account any schema overrides.
namepath_segment (list):
List of namepath codes defined in this object, not
including the container's namepath. (Implementation)
omit_ranks dict(str:int):
How many ranks to omit from the beginning in shortened
names for specific uses. (Implementation).
Probably a mistake this isn't in the FieldSchema instead.
fields (dict): The field values used to initialize the NameContext,
May include some details not defined in this attribution
API, and it includes the raw state of the 'name', 'code',
and 'title' fields, determining which are
authoritative -- i.e. fields which aren't specified are
left to 'float', being generated by the related ones.
Thus, name implies title, or title implies name. You
can have one or the other or both, but the other will be
generated from the provided one if it isn't specified.
(Implementation).
code (str|int|Enum):
Identification code for the unit (I replaced 'id' in earlier
versions, because 'id' is a reserved word in Python for Python
memory reference / pointer identity.
(R/W Property).
namepath (list):
List of codes for project units above this one.
(R/O Property, generated from namepath_segment and
container).
rank (int): Rank of this unit (same as in Schema).
(R/W Property, may affect other attributes).
name (str): Short name for the unit.
(R/W Property, may affect title).
title (str): Full title for the unit.
(R/W Property, may affect name).
designation (str):
Full designation for the unit, including all
the namepath elements, but no title.
fullname (str): The full designation, plus the current unit name.
shortname (str):Abbreviated designation, according to omit_ranks,
with name.
"""
def __init__(self, container, fields=None, namepath_segment=(), ):
self.clear()
if container or fields or namepath_segment:
self.update(container, fields, namepath_segment)
def clear(self):
self.fields = {}
self.schemas = ['project']
self.rank = 0
self.code = 'untitled'
self.container = None
self.namepath_segment = []
def update(self, container=None, fields=None, namepath_segment=()):
self.container = container
if namepath_segment:
self.namepath_segment = namepath_segment
else:
self.namepath_segment = []
try:
self.schemas = self.container.schemas
except AttributeError:
self.schemas = []
try:
self.omit_ranks = self.container.omit_ranks
except AttributeError:
self.omit_ranks = {}
self.omit_ranks.update({
'edit': 0,
'render': 1,
'filename': 1,
'scene': 3})
if fields:
if isinstance(fields, dict):
self.fields.update(fields)
elif isinstance(fields, str):
self.fields.update(yaml.safe_load(fields))
def update_fields(self, data):
self.fields.update(data)
def _load_schemas(self, schemas, start=0):
"""
Load schemas from a list of schema dictionaries.
@schemas: list of dictionaries containing schema field data (see FieldSchema).
The data will typically be extracted from YAML, and is
expected to be a list of dictionaries, each of which defines
fields understood by the FieldSchema class, to instantiate
FieldSchema objects. The result is a linked chain of schemas from
the top of the project tree down.
@start: if a start value is given, the top of the existing schema
chain is kept, and the provided schemas starts under the rank of
the start level in the existing schema. This is what happens when
the schema is locally overridden at some point in the hierarchy.
"""
self.schemas = self.schemas[:start]
if self.schemas:
last = self.schemas[-1]
else:
last = None
for schema in schemas:
self.schemas.append(FieldSchema(last, schema['rank'], schema=schema))
#last = self.schemas[-1]
def _parse_words(self, wordtext):
words = []
groups = re.split(r'[\W_]', wordtext)
for group in groups:
if len(group)>1:
group = group[0].upper() + group[1:]
words.extend(re.findall(r'[A-Z][a-z]*', group))
elif len(group)==1:
words.append(group[0].upper())
else:
continue
return words
def _cap_words(self, words):
return ''.join(w.capitalize() for w in words)
def _underlower_words(self, words):
return '_'.join(w.lower() for w in words)
def _undercap_words(self, words):
return '_'.join(w.capitalize() for w in words)
def _spacecap_words(self, words):
return ' '.join(w.capitalize() for w in words)
def _compress_name(self, name):
return self._cap_words(self._parse_words(name))
@property
def namepath(self):
if self.container:
return self.container.namepath + self.namepath_segment
else:
return self.namepath_segment
@property
def rank(self):
if 'rank' in self.fields:
return self.fields['rank']
else:
return None
@rank.setter
def rank(self, rank):
self.fields['rank'] = rank
@property
def name(self):
if 'name' in self.fields:
return self.fields['name']
elif 'title' in self.fields:
return self._compress_name(self.fields['title'])
# elif 'code' in self.fields:
# return self.fields['code']
else:
return ''
@name.setter
def name(self, name):
self.fields['name'] = name
@property
def code(self):
if self.rank:
return self.fields[self.rank]['code']
else:
return self.fields['code']
@code.setter
def code(self, code):
if self.rank:
self.fields[self.rank] = {'code': code}
else:
self.fields['code'] = code
@property
def description(self):
if 'description' in self.fields:
return self.fields['description']
else:
return ''
@description.setter
def description(self, description):
self.fields['description'] = str(description)
def _get_name_components(self):
components = []
for code, schema in zip(self.namepath, self.schemas):
if code is None: continue
components.append(schema.format.format(code))
components.append(schema.delimiter)
return components[:-1]
@property
def fullname(self):
if self.name:
return (self.designation +
self.schemas[-1].delimiter +
self._compress_name(self.name) )
else:
return self.designation
@property
def designation(self):
return ''.join(self._get_name_components())
@property
def shortname(self):
namebase = self.omit_ranks['filename']*2
return (''.join(self._get_name_components()[namebase:]) +
self.schemas[-1].delimiter +
self._compress_name(self.name))
def get_scene_name(self, suffix=''):
"""
Create a name for the current scene, based on namepath.
Arguments:
suffix (str): Optional suffix code used to improve
identifications of scenes.
"""
namebase = self.omit_ranks['scene']*2
desig = ''.join(self._get_name_components()[namebase:])
if suffix:
return desig + ' ' + suffix
else:
return desig
def get_render_path(self, suffix='', framedigits=5, ext='png'):
"""
Create a render filepath, based on namepath and parameters.
Arguments:
suffix (str):
Optional unique code (usually for render profile).
framedigits (int):
How many digits to reserve for frame number.
ext (str):
Filetype extension for the render.
This is meant to be called by render_profile to combine namepath
based name with parameters for the specific render, to uniquely
idenfify movie or image-stream output.
"""
desig = ''.join(self._get_name_components()[self.omit_ranks['render']+1:])
if ext in ('avi', 'mov', 'mp4', 'mkv'):
if suffix:
path = os.path.join(self.render_root, suffix,
desig + '-' + suffix + '.' + ext)
else:
path = os.path.join(self.render_root, ext.upper(),
desig + '.' + ext)
else:
if suffix:
path = os.path.join(self.render_root,
suffix, desig,
desig + '-' + suffix + '-f' + '#'*framedigits + '.' + ext)
else:
path = os.path.join(self.render_root,
ext.upper(), desig,
desig + '-f' + '#'*framedigits + '.' + ext)
return path

286
abx/name_schema.py Normal file
View File

@ -0,0 +1,286 @@
# name_schema.py
"""
Object for managing schema directives from project YAML and applying them to parsing and mapping name fields.
"""
import string, collections
from .ranks import RankNotFound, Rank, Branch, Trunk
class FieldSchema(object):
"""
Represents a schema used for parsing and constructing a field in names.
We need naming information in various formats, based on knowledge about
the role of the Blender file and scene in the project. This object tracks
this information and returns correct names based on it via properties.
Note that FieldSchema is NOT an individual project unit name, but a defined
pattern for how names are treat at that level in the project. It is a class
of names, not a name itself. Thus "shot" has a schema, but is distinct from
"shot A" which is a particular "project unit". The job of the schema is
to tell us things like "shots in this project will be represented by
single capital letters".
See NameContext for the characteristics of a particular unit.
Attributes:
codetype (type): Type of code name used for this rank.
Usually it will be int, str, or Enum.
Pre-defined enumerations are available
for uppercase letters (_letters) and
lowercase letters (_lowercase) (Roman --
in principle, other alphabets could be added).
rank (Rank): Rank of hierarchy under project (which is 0). The
rank value increases as you go "down" the tree.
Sorry about that confusion.
ranks (Branch): List of named ranks known to schema (may include
both higher and lower ranks).
parent (|None):
Earlier rank to which this schema is attached.
format (str): Code for formatting with Python str.format() method.
Optional: format can also be specified with the
following settings, or left to default formatting.
pad (str): Padding character.
minlength (int): Minimum character length (0 means it may be empty).
maxlength (int): Maximum character length (0 means no limit).
words (bool): Treat name/title fields like a collection of words,
which can then be represented using "TitleCaps" or
"underscore_spacing", etc for identifier use.
delimiter (str): Field delimiter marking the end of this ranks'
code in designations. Note this is the delimiter
after this rank - the higher (lower value) rank
controls the delimiter used before it.
default: The default value for this rank. May be None,
in which case, the rank will be treated as unset
until a setting is made. The UI must provide a
means to restore the unset value. Having no values
set below a certain rank is how a NameContext's
rank is determined.
Note that the rank may go back to a lower value than the schema's
parent object in order to overwrite earlier schemas (for overriding a
particular branch in the project) or (compare this to the use of '..'
in operating system paths). Or it may skip a rank, indicating an
implied intermediate value, which will be treated as having a fixed
value. (I'm not certain I want that, but it would allow us to keep
rank numbers synchronized better in parallel hierarchies in a project).
Note that schemas can be overridden at any level in a project by
'project_schema' directives in unit YAML files, so it is possible to
change the schema behavior locally. By design, only lower levels in
the hierarchy (higher values of rank) can be affected by overrides.
This kind of use isn't fully developed yet, but the plan is to include
things like managing 'Library' assets with a very different structure
from shot files. This way, the project can split into 'Library' and
'Episode' forks with completely different schemas for each.
"""
# Defaults
_default_schema = {
'delimiter':'-',
'type': 'string',
'format':'{:s}',
'minlength':1, # Must be at least one character
'maxlength':0, # 0 means unlimited
'words': False, # If true, treat value as words and spaces
'pad': '0', # Left-padding character for fixed length
'default': None,
'rank': 'project',
'irank': 0,
'ranks': ('series', 'episode', 'sequence',
'block', 'camera', 'shot', 'element')
}
# Really this is more like a set than a dictionary right now, but I
# thought I might refactor to move the definitions into the dictionary:
_codetypes = {
'number':{},
'string':{},
'letter':{},
'lowercase':{},
}
_letter = tuple((A,A,A) for A in string.ascii_uppercase)
_lowercase = tuple((a,a,a) for a in string.ascii_lowercase)
rank = 'project'
irank = 0
default = None
#ranks = ('project',)
branch = Trunk
def __init__(self, parent=None, rank=None, schema=None, debug=False):
"""
Create a FieldSchema from schema data source.
FieldSchema is typically initialized based on data from YAML files
within the project. This allows us to avoid encoding project structure
into ABX, leaving how units are named up to the production designer.
If you want our suggestions, you can look at the "Lunatics!" project's
'lunatics.yaml' file, or the 'myproject.yaml' file in the ABX source
distribution.
Arguments:
parent (FieldSchema):
The level in the schema hierarchy above this one.
Should be None if this is the top.
rank (int): The rank of this schema to be created.
schema (dict): Data defining the schema, typically loaded from a
YAML file in the project.
debug (bool): Used only for testing. Turns on some verbose output
about internal implementation.
Note that the 'rank' is specified because it may NOT be sequential from
the parent schema.
"""
# Three types of schema data:
# Make sure schema is a copy -- no side effects!
if not schema:
schema = {}
else:
s = {}
s.update(schema)
schema = s
if not rank and 'rank' in schema:
rank = schema['rank']
# Stepped down in rank from parent:
self.parent = parent
if parent and rank:
# Check rank is defined in parent ranks and use that
# We can skip optional ranks
if rank in parent.ranks:
j = parent.ranks.index(rank)
self.ranks = parent.ranks[j+1:]
self.rank = rank
else:
# It's an error to ask for a rank that isn't defined
raise RankNotFound(
'"%s" not in defined ranks for "%s"' % (rank, parent))
elif parent and not rank:
# By default, get the first rank below parent
self.rank = parent.ranks[0]
self.ranks = parent.ranks[1:]
elif rank and not parent:
# With no parent, we're starting a new tree and renaming the root
self.rank = rank
self.ranks = self._default_schema['ranks']
else: # not rank and not parent:
# New tree with default rank
self.rank = self._default_schema['rank']
self.ranks = self._default_schema['ranks']
# Directly inherited/acquired from parent
# So far, only need a delimiter specified, but might be other stuff
self.delimiter = self._default_schema['delimiter']
if parent and parent.delimiter: self.delimiter = parent.delimiter
# Explicit override by the new schema:
if 'ranks' in schema: self.ranks = schema['ranks']
if 'delimiter' in schema: self.delimiter = schema['delimiter']
if 'default' in schema:
if schema['default'] == 'None':
self.default = None
else:
self.default = schema['default']
# Default unless specified (i.e. not inherited from parent)
newschema = {}
newschema.update(self._default_schema)
newschema.update(schema)
self.format = str(newschema['format'])
self.minlength = int(newschema['minlength'])
self.maxlength = int(newschema['maxlength'])
self.pad = str(newschema['pad'])
self.words = bool(newschema['words'])
if newschema['type'] == 'letter':
self.codetype = self._letter
elif newschema['type'] == 'lowercase':
self.codetype = self._lowercase
elif newschema['type'] == 'number':
# Recognized Python types
self.codetype = int
if 'minlength' or 'maxlength' in schema:
self.format = '{:0>%dd}' % self.minlength
elif newschema['type'] == 'string':
self.codetype = str
if ('minlength' in schema) or ('maxlength' in schema):
if self.maxlength == 0:
# Special case for unlimited length
self.format = '{:%1.1s>%ds}' % (self.pad, self.minlength)
self.format = '{:%1.1s>%d.%ds}' % (
self. pad, self.minlength, self.maxlength)
elif newschema['type'] == 'bool':
self.codetype = bool
elif isinstance(newschema['type'], collections.Sequence):
# Enumerated types
# This is somewhat specific to Blender -- setting the
# enumeration values requires a sequence in a particular format
self.codetype = []
for option in newschema['type']:
if type(option) is not str and isinstance(option, collections.Sequence):
option = tuple([str(e) for e in option][:3])
else:
option = (str(option), str(option), str(option))
self.codetype.append(option)
elif isinstance(newschema['type'], collections.Mapping):
self.codetype = []
for key, val in newschema['type'].items():
if type(val) is not str and isinstance(val, collections.Sequence):
if len(val) == 0:
option = (str(key), str(key), str(key))
elif len(val) == 1:
option = (str(key), str(val[0]), str(val[0]))
else:
option = (str(key), str(val[0]), str(val[1]))
else:
option = (str(key), str(val), str(val))
self.codetype.append(option)
else:
# If all else fails
self.codetype = None
def __repr__(self):
return('<(%s).FieldSchema: %s (%s, %s, %s, (%s))>' % (
repr(self.parent),
#self.irank,
self.rank,
self.delimiter,
self.default,
self.format,
self.codetype
))

16
abx/parsers/__init__.py Normal file
View File

@ -0,0 +1,16 @@
# parsers (sub-package)
"""
Filename Parsers & Registry for FileContext.
"""
NameParsers = {} # Parser registry
def registered_parser(parser):
"""
Decorator function to register a parser class.
"""
NameParsers[parser.name] = parser
return parser
from . import abx_episode, abx_fallback, abx_schema

Binary file not shown.

Binary file not shown.

Binary file not shown.

Binary file not shown.

204
abx/parsers/abx_episode.py Normal file
View File

@ -0,0 +1,204 @@
# abx_episode.py
"""
Custom parser written for "Lunatics!" Project Episode files.
Superseded by 'abx_schema' parser (probably).
"""
import re, copy
from . import registered_parser
wordre = re.compile(r'([A-Z][a-z]+|[a-z]+|[0-9]+|[A-Z][A-Z]+)')
@registered_parser
class Parser_ABX_Episode:
"""
Original "Lunatics!" filename parsing algorithm. (DEPRECATED)
This parser was written before the Schema parser. It hard-codes the schema used
in the "Lunatics!" Project, and can probably be safely replaced by using the Schema
parser with appropriate YAML settings in the <project>.yaml file, which also allows
much more flexibility in naming schemes.
YAML parameter settings available for this parser:
---
definitions:
parser: abx_episode # Force use of this parser
parser_options: # Available settings (w/ defaults)
field_separator: '-'
episode_separator: 'E'
filetype_separator: '.'
Filetypes and roles are hard-code, and can't be changed from the YAML.
Assumes field-based filenames of the form:
<series>E<episode>[-<seq>[-<block>[-Cam<camera>][-<shot>]]][-<title>]-<role>.<filetype>
Where the <field> indicates fields with fieldnames, and there are three expected separators:
- is the 'field_separator'
E is the 'episode_separator'
. is the 'filetype_separator'
(These can be overridden in the initialization).
The class is callable, taking a string as input and returning a dictionary of fields.
"""
name = 'abx_episode'
max_score = 10 # Maximum number of fields parsed
# supported values for filetype
filetypes = {
'blend': "Blender File",
'kdenlive': "Kdenlive Video Editor File",
'mlt': "Kdenlive Video Mix Script",
'svg': "Scalable Vector Graphics (Inkscape)",
'kra': "Krita Graphic File",
'xcf': "Gimp Graphic File",
'png': "Portable Network Graphics (PNG) Image",
'jpg': "Joint Photographic Experts Group (JPEG) Image",
'aup': "Audacity Project",
'ardour': "Ardour Project",
'flac': "Free Lossless Audio Codec (FLAC)",
'mp3': "MPEG Audio Layer III (MP3) Audio File",
'ogg': "Ogg Vorbis Audio File",
'avi': "Audio Video Interleave (AVI) Video Container",
'mkv': "Matroska Video Container",
'mp4': "Moving Picture Experts Group (MPEG) 4 Format}",
'txt': "Plain Text File"
}
# Roles that make sense in an episode context
roles = {
'extras': "Extras, crowds, auxillary animated movement",
'mech': "Mechanical animation",
'anim': "Character animation",
'cam': "Camera direction",
'vfx': "Visual special effects",
'compos': "Compositing",
'bkg': "Background 2D image",
'bb': "Billboard 2D image",
'tex': "Texture 2D image",
'foley': "Foley sound",
'voice': "Voice recording",
'fx': "Sound effects",
'music': "Music track",
'cue': "Musical cue",
'amb': "Ambient sound",
'loop': "Ambient sound loop",
'edit': "Video edit"
}
# A few filetypes imply their roles:
roles_by_filetype = {
'kdenlive': 'edit',
'mlt': 'edit'
}
def __init__(self, field_separator='-', episode_separator='E', filetype_separator='.',
fields=None, filetypes=None, roles=None, **kwargs):
if not fields:
fields = {}
if filetypes:
self.filetypes = copy.deepcopy(self.filetypes) # Copy class attribute to instance
self.filetypes.update(filetypes) # Update with new values
if roles:
self.roles = copy.deepcopy(self.roles) # Copy class attribute to instance
self.roles.update(roles) # Update with new values
self.field_separator = field_separator
self.episode_separator = episode_separator
self.filetype_separator = filetype_separator
def __call__(self, filename, namepath):
score = 0.0
fielddata = {}
# Check for filetype ending
i_filetype = filename.rfind(self.filetype_separator)
if i_filetype < 0:
fielddata['filetype'] = None
else:
fielddata['filetype'] = filename[i_filetype+1:]
filename = filename[:i_filetype]
score = score + 1.0
components = filename.split(self.field_separator)
# Check for role marker in last component
if components[-1] in self.roles:
fielddata['role'] = components[-1]
del components[-1]
fielddata['hierarchy'] = 'episode'
score = score + 2.0
elif fielddata['filetype'] in self.roles_by_filetype:
fielddata['role'] = self.roles_by_filetype[fielddata['filetype']]
fielddata['hierarchy'] = 'episode'
else:
fielddata['role'] = None
fielddata['hierarchy'] = None
# Check for a descriptive title (must be 3+ characters in length)
if components and len(components[-1])>2:
# Normalize the title as words with spaces
title = ' '.join(w for w in wordre.split(components[-1]) if wordre.fullmatch(w))
del components[-1]
score = score + 1.0
else:
title = None
# Check if first field contains series/episode number
if components:
prefix = components[0]
try:
fielddata['series'] = {}
fielddata['episode'] = {}
fielddata['series']['code'], episode_id = prefix.split(self.episode_separator)
fielddata['episode']['code'] = int(episode_id)
fielddata['rank'] = 'episode'
del components[0]
score = score + 2.0
except:
pass
# Check for sequence/block/shot/camera designations
if components:
fielddata['seq'] = {}
fielddata['seq']['code'] = components[0]
fielddata['rank'] = 'seq'
del components[0]
score = score + 1.0
if components:
try:
fielddata['block'] = {}
fielddata['block']['code'] = int(components[0])
del components[0]
fielddata['rank'] = 'block'
score = score + 1.0
except:
pass
if components and components[0].startswith('Cam'):
fielddata['camera'] = {}
fielddata['camera']['code'] = components[0][len('Cam'):]
fielddata['rank'] = 'camera'
del components[0]
score = score + 1.0
if components:
# Any remaining structure is joined back to make the shot ID
fielddata['shot'] = {}
fielddata['shot']['code'] = ''.join(components)
fielddata['rank'] = 'shot'
components = None
score = score + 1.0
if title and fielddata['rank'] in fielddata:
fielddata[fielddata['rank']]['title'] = title
return score/self.max_score, fielddata

105
abx/parsers/abx_fallback.py Normal file
View File

@ -0,0 +1,105 @@
# abx_fallback.py
"""
Fallback parser used in case others fail.
The fallback parser makes only a very minimal and robust set of assumptions.
Any legal filename will successfully return a simple parse, though much
interpretation may be lost. It still allows for common field-based practices,
but falls back on using the unaltered filename if necessary.
"""
import re, os
import yaml
from . import registered_parser
DEFAULT_YAML = {}
with open(os.path.join(os.path.dirname(__file__), '..', 'abx.yaml')) as def_yaml_file:
DEFAULT_YAML.update(yaml.safe_load(def_yaml_file))
@registered_parser
class Parser_ABX_Fallback(object):
"""
Highly-tolerant parser to fall back on if others fail.
Makes very minimal assumptions about filename structure.
YAML options available:
---
definitions:
parser: abx_fallback # Force use of this parser.
There are no other options. Field separators are defined very broadly,
and include most non-word characters (~#$!=+&_-). This was mostly designed
to work without a project schema available.
"""
name = 'abx_fallback'
filetypes = DEFAULT_YAML['definitions']['filetypes']
roles = DEFAULT_YAML['definitions']['roles']
roles_by_filetype = (
DEFAULT_YAML['definitions']['roles_by_filetype'])
main_sep_re = re.compile(r'\W+') # Any single non-word char
comment_sep_re = re.compile(r'[\W_][\W_]+|[~#$!=+&]+')
def __init__(self, **kwargs):
pass
def _parse_ending(self, filename, separator):
try:
remainder, suffix = filename.rsplit(separator, 1)
score = 1.0
except ValueError:
remainder = filename
suffix = None
score = 0.0
return (suffix, remainder, score)
def __call__(self, filename, namepath):
fields = {}
score = 1.0
possible = 4.5
split = filename.rsplit('.', 1)
if len(split)<2 or split[1] not in self.filetypes:
fields['filetype'] = None
remainder = filename
score += 1.0
else:
fields['filetype'] = split[1]
remainder = split[0]
comment_match = self.comment_sep_re.search(remainder)
if comment_match:
fields['comment'] = remainder[comment_match.end():]
remainder = remainder[:comment_match.start()]
else:
fields['comment'] = None
role = self.main_sep_re.split(remainder)[-1]
if role in self.roles:
fields['role'] = role
remainder = remainder[:-1-len(role)]
score += 1.0
else:
fields['role'] = None
# Implied role
if fields['filetype'] in self.roles_by_filetype:
fields['role'] = self.roles_by_filetype[fields['filetype']]
score += 1.0
words = self.main_sep_re.split(remainder)
fields['code'] = ''.join([w.capitalize() for w in words])
fields['title'] = remainder
return score/possible, fields

189
abx/parsers/abx_schema.py Normal file
View File

@ -0,0 +1,189 @@
# abx_schema.py
"""
Generalized fields-based parser based on provided schema.
Expands on the 'abx_episode' parser by allowing all the schema to
be defined by outside configuration data (generally provided in a
project YAML file, but this module does not depend on the data
source used).
"""
from . import registered_parser
@registered_parser
class Parser_ABX_Schema(object):
"""
Parser based on using the list of schemas.
The schemas are normally defined in the project root directory YAML.
The project YAML can additionally control parsing with this parser:
---
definitions:
parser: abx_schema # Force use of this parser
parser_options: # Set parameters
filetype_separator: '.'
comment_separator: '--'
role_separator: '-'
title_separator: '-'
filetypes: # Recognized filetypes.
blend: Blender File # <filetype>: documentation
...
roles: # Recognized role fields.
anim: Character Animation # <role>: documentation
...
roles_by_filetype: # Roles implied by filetype.
kdenlive: edit # <filetype>:<role>
...
(For the full default lists see abx/abx.yaml).
schemas (list): The current schema-list defining how filenames should be parsed.
This "Schema" parser uses this to determine both parsing and
mapping of text fields in the filename.
definitions(dict): The project definitions currently visible to the parser.
"""
name = 'abx_schema'
def __init__(self, schemas=None, definitions=None,
filetype_separator = '.',
comment_separator = '--',
role_separator = '-',
title_separator = '-',
**kwargs):
self.filetype_separator = filetype_separator
self.comment_separator = comment_separator
self.role_separator = role_separator
self.title_separator = title_separator
self.schemas = schemas
if 'roles' in definitions:
self.roles = definitions['roles']
else:
self.roles = []
if 'filetypes' in definitions:
self.filetypes = definitions['filetypes']
else:
self.filetypes = []
if 'roles_by_filetype' in definitions:
self.roles_by_filetype = definitions['roles_by_filetype']
else:
self.roles_by_filetype = []
def _parse_ending(self, filename, separator):
try:
remainder, suffix = filename.rsplit(separator, 1)
score = 1.0
except ValueError:
remainder = filename
suffix = None
score = 0.0
return (suffix, remainder, score)
def _parse_beginning(self, filename, separator):
try:
prefix, remainder = filename.split(separator, 1)
score = 1.0
except ValueError:
prefix = filename
remainder = ''
score = 0.0
return (prefix, remainder, score)
def __call__ (self, filename, namepath, debug=False):
fields = {}
score = 0.0
possible = 0.0
# First get specially-handled extensions
remainder = filename
field, newremainder, s = self._parse_ending(remainder, self.filetype_separator)
if field and field in self.filetypes:
remainder = newremainder
fields['filetype'] = field
score += s*1.0
else:
fields['filetype'] = None
field, remainder, s = self._parse_ending(remainder, self.comment_separator)
fields['comment'] = field
score += s*0.5
field, newremainder, s = self._parse_ending(remainder, self.role_separator)
if field and field in self.roles:
remainder = newremainder
fields['role'] = field
score += s*0.5
else:
fields['role'] = None
field, remainder, s = self._parse_ending(remainder, self.title_separator)
fields['title'] = field
score += s*0.5
possible += 3.0
# Implicit roles
if ( not fields['role'] and
fields['filetype'] and
fields['role'] in self.roles_by_filetype):
self.role = self.roles_by_filetype[fields['filetype']]
score += 0.2
#possible += 0.2
# Figure the rest out from the schema
# Find the matching rank start position for the filename
start = 0
for start, (schema, name) in enumerate(zip(self.schemas, namepath)):
field, r, s = self._parse_beginning(remainder, schema.delimiter)
try:
if field.lower() == schema.format.format(name).lower():
score += 1.0
break
except ValueError:
print(' (365) field, format', field, schema.format)
possible += 1.0
# Starting from that position, try to match fields
# up to the end of the namepath (checking against it)
irank = 0
for irank, (schema, name) in enumerate(
zip(self.schemas[start:], namepath[start:])):
if not remainder: break
field, remainder, s = self._parse_beginning(remainder, schema.delimiter)
score += s
try:
if ( type(field) == str and
field.lower() == schema.format.format(name).lower()):
fields[schema.rank]={'code':field}
fields['rank'] = schema.rank
score += 1.0
except ValueError:
print(' (384) field, format', field, schema.format)
possible += 2.0
# Remaining fields are authoritative (doesn't affect score)
for schema in self.schemas[irank:]:
if not remainder: break
field, remainder, s = self._parse_beginning(remainder, schema.delimiter)
fields[schema.rank]={'code':field}
fields['rank'] = schema.rank
if 'rank' in fields:
fields[fields['rank']]['title'] = fields['title']
if not fields['role'] and fields['filetype'] in self.roles_by_filetype:
fields['role'] = self.roles_by_filetype[fields['filetype']]
return score/possible, fields

View File

@ -7,7 +7,11 @@ possibility of branching at nodes with redefined ranks via
the 'project_schema' directives in project YAML files. the 'project_schema' directives in project YAML files.
""" """
import numbers class RankNotFound(LookupError):
"""
Error returned if an unexpected 'rank' is encountered.
"""
pass
class Branch(object): class Branch(object):
""" """
@ -37,7 +41,7 @@ class Branch(object):
if self.code: if self.code:
code = self.code code = self.code
else: else:
code = 'trunk' code = 'Trunk'
return "<branch '%s': %s>" % (code, ranklist) return "<branch '%s': %s>" % (code, ranklist)
def __contains__(self, other): def __contains__(self, other):
@ -53,9 +57,13 @@ class Branch(object):
if isinstance(n, int) and 0 < n < len(self._ranks): if isinstance(n, int) and 0 < n < len(self._ranks):
return self._ranks[n] return self._ranks[n]
elif isinstance(n, str): elif isinstance(n, str):
if n.lower()=='trunk':
return self._ranks[0]
for rank in self._ranks: for rank in self._ranks:
if str(rank) == n: if str(rank) == n:
return rank return rank
elif n==0:
self._ranks[0]
else: else:
raise TypeError raise TypeError
@ -168,7 +176,7 @@ class Rank(object):
if (self.num + other) < len(self.branch.ranks): if (self.num + other) < len(self.branch.ranks):
return self.branch.ranks[self.num+other] return self.branch.ranks[self.num+other]
elif (self.num + other) < 1: elif (self.num + other) < 1:
return trunk return Trunk
else: else:
return None return None
else: else:
@ -184,7 +192,7 @@ class Rank(object):
if 0 < (self.num - other) < len(self.branch.ranks): if 0 < (self.num - other) < len(self.branch.ranks):
return self.branch.ranks[self.num-other] return self.branch.ranks[self.num-other]
elif (self.num - other) < 1: elif (self.num - other) < 1:
return trunk return Trunk
elif (self.num - other) > len(self.branch.ranks): elif (self.num - other) > len(self.branch.ranks):
return None return None
else: else:
@ -241,7 +249,7 @@ class RankList(list):
return super().__getitem__(i) return super().__getitem__(i)
# Define the trunk branch object # Define the Trunk branch object
# This schema will make sense for any unaffiliated Blender document, # This schema will make sense for any unaffiliated Blender document,
# even if it hasn't been saved as a file yet: # even if it hasn't been saved as a file yet:
trunk = Branch(None, '', 0, ('', 'file', 'scene')) Trunk = Branch(None, '', 0, ('', 'file', 'scene'))

View File

@ -521,7 +521,7 @@ class Parser_ABX_Fallback(object):
class RankNotFound(LookupError): class RankNotFound(LookupError):
pass pass
class NameSchema(object): class FieldSchema(object):
""" """
Represents a schema used for parsing and constructing designations, names, etc. Represents a schema used for parsing and constructing designations, names, etc.
""" """
@ -686,7 +686,7 @@ class NameSchema(object):
def __repr__(self): def __repr__(self):
return('<(%s).NameSchema: %s (%s, %s, %s, (%s))>' % ( return('<(%s).FieldSchema: %s (%s, %s, %s, (%s))>' % (
repr(self.parent), repr(self.parent),
#self.irank, #self.irank,
self.rank, self.rank,
@ -751,11 +751,11 @@ class NameContext(object):
""" """
Load schemas from a list of schema dictionaries. Load schemas from a list of schema dictionaries.
@schemas: list of dictionaries containing schema field data (see NameSchema). @schemas: list of dictionaries containing schema field data (see FieldSchema).
The data will typically be extracted from YAML, and is The data will typically be extracted from YAML, and is
expected to be a list of dictionaries, each of which defines expected to be a list of dictionaries, each of which defines
fields understood by the NameSchema class, to instantiate fields understood by the FieldSchema class, to instantiate
NameSchema objects. The result is a linked chain of schemas from FieldSchema objects. The result is a linked chain of schemas from
the top of the project tree down. the top of the project tree down.
@start: if a start value is given, the top of the existing schema @start: if a start value is given, the top of the existing schema
@ -769,7 +769,7 @@ class NameContext(object):
else: else:
last = None last = None
for schema in schemas: for schema in schemas:
self.schemas.append(NameSchema(last, schema['rank'], schema=schema)) self.schemas.append(FieldSchema(last, schema['rank'], schema=schema))
#last = self.schemas[-1] #last = self.schemas[-1]
def _parse_words(self, wordtext): def _parse_words(self, wordtext):

View File

@ -63,9 +63,9 @@ class FileContext_Utilities_Tests(unittest.TestCase):
class FileContext_NameSchema_Interface_Tests(unittest.TestCase): class FileContext_NameSchema_Interface_Tests(unittest.TestCase):
""" """
Test the interfaces presented by NameSchema. Test the interfaces presented by FieldSchema.
NameSchema is not really intended to be used from outside the FieldSchema is not really intended to be used from outside the
file_context module, but it is critical to the behavior of the file_context module, but it is critical to the behavior of the
module, so I want to make sure it's working as expected. module, so I want to make sure it's working as expected.
""" """
@ -89,58 +89,6 @@ class FileContext_NameSchema_Interface_Tests(unittest.TestCase):
{'rank': 'shot', 'delimiter':'-', 'format':'{!s:s}'}, {'rank': 'shot', 'delimiter':'-', 'format':'{!s:s}'},
{'rank': 'element', 'delimiter':'-', 'format':'{!s:s}'}] {'rank': 'element', 'delimiter':'-', 'format':'{!s:s}'}]
def test_NameSchema_create_single(self):
ns = file_context.NameSchema(schema = self.TESTSCHEMA_LIST[0])
# Test for ALL the expected properties:
# Set by the test schema
self.assertEqual(ns.rank, 'project')
self.assertEqual(ns.delimiter, '-')
self.assertEqual(ns.format, '{:s}')
self.assertEqual(ns.words, True)
self.assertEqual(ns.codetype, str)
# Default values
self.assertEqual(ns.pad, '0')
self.assertEqual(ns.minlength, 1)
self.assertEqual(ns.maxlength, 0)
self.assertEqual(ns.default, None)
# Candidates for removal:
self.assertEqual(ns.irank, 0) # Is this used at all?
self.assertEqual(ns.parent, None)
self.assertListEqual(list(ns.ranks),
['series', 'episode', 'sequence',
'block', 'camera', 'shot', 'element'])
def test_NameSchema_load_chain_from_project_yaml(self):
with open(self.TESTPROJECTYAML, 'rt') as yaml_file:
data = yaml.safe_load(yaml_file)
schema_dicts = data['project_schema']
schema_chain = []
last = None
for schema_dict in schema_dicts:
rank = schema_dict['rank']
parent = last
schema_chain.append(file_context.NameSchema(
parent = parent,
rank = rank,
schema = schema_dict))
last = schema_chain[-1]
#print( schema_chain )
self.assertEqual(len(schema_chain), 8)
self.assertEqual(
schema_chain[-1].parent.parent.parent.parent.parent.parent.parent.rank,
'project')
self.assertEqual(schema_chain[5].rank, 'camera')
self.assertEqual(schema_chain[5].codetype[1], ('c2', 'c2', 'c2'))
@ -218,7 +166,7 @@ class FileContext_Parser_UnitTests(unittest.TestCase):
'A.001-LP-1-BeginningOfEnd-anim.txt') 'A.001-LP-1-BeginningOfEnd-anim.txt')
def setUp(self): def setUp(self):
self.TESTSCHEMAS = [file_context.NameSchema( #rank=s['rank'], self.TESTSCHEMAS = [file_context.FieldSchema( #rank=s['rank'],
schema=s) schema=s)
for s in self.TESTSCHEMA_LIST] for s in self.TESTSCHEMA_LIST]

100
tests/test_name_schema.py Normal file
View File

@ -0,0 +1,100 @@
#!/usr/bin/env python3
"""
Test the file_context module.
This was written well after I wrote the module, and starts out as a conversion
from the doctests I had in the module already.
"""
import unittest, os
import yaml
import sys
print("__file__ = ", __file__)
sys.path.append(os.path.normpath(os.path.join(__file__, '..', '..')))
from abx import name_schema
class FileContext_NameSchema_Interface_Tests(unittest.TestCase):
"""
Test the interfaces presented by FieldSchema.
FieldSchema is not really intended to be used from outside the
file_context module, but it is critical to the behavior of the
module, so I want to make sure it's working as expected.
"""
TESTDATA = os.path.abspath(
os.path.join(os.path.dirname(__file__), '..', 'testdata'))
TESTPROJECTYAML = os.path.join(TESTDATA, 'myproject', 'myproject.yaml')
TESTPATH = os.path.join(TESTDATA, 'myproject/Episodes/' +
'A.001-Pilot/Seq/LP-LastPoint/' +
'A.001-LP-1-BeginningOfEnd-anim.txt')
# Normally from 'project_schema' in YAML
TESTSCHEMA_LIST =[
{'rank': 'project', 'delimiter':'-', 'format':'{:s}', 'words':True},
{'rank': 'series', 'delimiter':'E', 'format':'{:2s}'},
{'rank': 'episode', 'delimiter':'-', 'format':'{!s:>02s}'},
{'rank': 'sequence','delimiter':'-', 'format':'{:2s}'},
{'rank': 'block', 'delimiter':'-', 'format':'{!s:1s}'},
{'rank': 'shot', 'delimiter':'-', 'format':'{!s:s}'},
{'rank': 'element', 'delimiter':'-', 'format':'{!s:s}'}]
def test_NameSchema_create_single(self):
ns = name_schema.FieldSchema(schema = self.TESTSCHEMA_LIST[0])
# Test for ALL the expected properties:
# Set by the test schema
self.assertEqual(ns.rank, 'project')
self.assertEqual(ns.delimiter, '-')
self.assertEqual(ns.format, '{:s}')
self.assertEqual(ns.words, True)
self.assertEqual(ns.codetype, str)
# Default values
self.assertEqual(ns.pad, '0')
self.assertEqual(ns.minlength, 1)
self.assertEqual(ns.maxlength, 0)
self.assertEqual(ns.default, None)
# Candidates for removal:
self.assertEqual(ns.irank, 0) # Is this used at all?
self.assertEqual(ns.parent, None)
self.assertListEqual(list(ns.ranks),
['series', 'episode', 'sequence',
'block', 'camera', 'shot', 'element'])
def test_NameSchema_load_chain_from_project_yaml(self):
with open(self.TESTPROJECTYAML, 'rt') as yaml_file:
data = yaml.safe_load(yaml_file)
schema_dicts = data['project_schema']
schema_chain = []
last = None
for schema_dict in schema_dicts:
rank = schema_dict['rank']
parent = last
schema_chain.append(name_schema.FieldSchema(
parent = parent,
rank = rank,
schema = schema_dict))
last = schema_chain[-1]
#print( schema_chain )
self.assertEqual(len(schema_chain), 8)
self.assertEqual(
schema_chain[-1].parent.parent.parent.parent.parent.parent.parent.rank,
'project')
self.assertEqual(schema_chain[5].rank, 'camera')
self.assertEqual(schema_chain[5].codetype[1], ('c2', 'c2', 'c2'))

View File

@ -12,17 +12,17 @@ from abx import ranks
class BranchTests(unittest.TestCase): class BranchTests(unittest.TestCase):
def test_trunk_branch(self): def test_trunk_branch(self):
t = ranks.trunk.rank('') t = ranks.Trunk.rank('')
f = ranks.trunk.rank('file') f = ranks.Trunk.rank('file')
s = ranks.trunk.rank('scene') s = ranks.Trunk.rank('scene')
self.assertEqual(repr(ranks.trunk), "<branch 'trunk': file, scene>") self.assertEqual(repr(ranks.Trunk), "<branch 'Trunk': file, scene>")
self.assertIn(t, ranks.trunk) self.assertIn(t, ranks.Trunk)
self.assertIn(f, ranks.trunk) self.assertIn(f, ranks.Trunk)
self.assertIn(s, ranks.trunk) self.assertIn(s, ranks.Trunk)
def test_defining_branch(self): def test_defining_branch(self):
b = ranks.Branch(ranks.trunk, 'myproject', 1, b = ranks.Branch(ranks.Trunk, 'myproject', 1,
('project', 'series', 'episode', 'sequence', ('project', 'series', 'episode', 'sequence',
'block', 'shot', 'element')) 'block', 'shot', 'element'))
@ -30,7 +30,7 @@ class BranchTests(unittest.TestCase):
class RanksTests(unittest.TestCase): class RanksTests(unittest.TestCase):
def setUp(self): def setUp(self):
self.b = ranks.Branch(ranks.trunk, 'myproject', 1, self.b = ranks.Branch(ranks.Trunk, 'myproject', 1,
('project', 'series', 'episode', 'sequence', ('project', 'series', 'episode', 'sequence',
'block', 'shot', 'element')) 'block', 'shot', 'element'))
@ -115,12 +115,12 @@ class RanksTests(unittest.TestCase):
pr = self.b.rank('project') pr = self.b.rank('project')
r = se - 1 # Normal - 'project' is one below 'series' r = se - 1 # Normal - 'project' is one below 'series'
s = se - 2 # ? Should this be 'project' or 'trunk'/None? s = se - 2 # ? Should this be 'project' or 'Trunk'/None?
t = se - 3 # "` " t = se - 3 # "` "
self.assertEqual(r, pr) self.assertEqual(r, pr)
self.assertEqual(s, ranks.trunk) self.assertEqual(s, ranks.Trunk)
self.assertEqual(t, ranks.trunk) self.assertEqual(t, ranks.Trunk)
def test_rank_slices_from_branch(self): def test_rank_slices_from_branch(self):
@ -156,4 +156,14 @@ class RanksTests(unittest.TestCase):
ranks.branch, ranks.branch,
self.b) self.b)
def test_using_rank_as_key(self):
d = dict(zip(self.b.ranks, range(len(self.b.ranks))))
R = self.b.rank
self.assertDictEqual(d, {
R('trunk'):0, R('project'):1, R('series'):2, R('episode'):3,
R('sequence'):4, R('block'):5, R('shot'):6, R('element'):7
})