autotagical

autotagical is a utility to automagically rename and sort tagged files (such as those produced by TagSpaces) according to user-defined schemas. It reads in tagged files from one of more input directories then renames and/or moves them to an output folder hierarchy according to rules specified in user-provided schemas. It is intended for use in concert with file tagging software, e.g. TagSpaces.

Getting Started/Installation

autotagical may be most easily installed with pip by running:

pip install autotagical

If you'd like to run autotagical by cloning this repository, then you'll need to install the following requirements, e.g. with pip:

• setuptools
• jsonschema>=3
• packaging

Using autotagical

Usage

autotagical [-h] [-V] [-C <config file>] [-H] [-i <input path>]
            [-I <ignore file>] [-R] [-o <output path>] [-O]
            [-g <tag group file>] [-s <schema file>] [-A] [--cleanin]
            [--cleanout] [-c] [-F] [-k] [-m] [-M] [-n] [-N] [-t]
            [--debug] [-l <log file>] [-L] [-P] [-q] [-v] [--force]
            [--yes]

Help Options

These options display helpful information and exit.

• [-h/--help] -- Display a help/usage message and exit.
• [-V/--version] -- Display the current version and information about known file formats and exit.

Configuration Options

• [-C/--config <config file>] -- Loads the config file at the specified path.

Input Options

These options determine behavior loading in files to be moved and/or renamed. At least one input folder must be specified.

• [-H/--hidden] -- Process hidden files (and directories, if -R is specified).
• [-i/--input <input folder>] -- Path to a folder with input files. May be specified more than once.
• [-I/--ignore <ignore file>] -- Path to file patterns (regex format) to ignore (each on new line). May be specified more than once.
• [-R/--recursive] -- Load files recursively from input folders, i.e. descend into subfolders.

Output Options

These options determine behavior in outputting move/renamed files (but do not specify the rules by which they are to be moved and/or renamed). At least one output folder must be specified (by one option or the other).

• [-o/--output] -- Path to a root folder to output files to. May be specified more than once (output will be duplicated to each).
• [-O/--organize] -- Organize files in place (i.e. use the first input folder for output). Often used with -R.

Schema Options

These options specify the rules for moving/renaming files and are the heart of autotagical. Information on the structure of these files may be found below in the Tag Group Format and Schema Format sections.

• [-g/--groups <tag group file>] -- Path to a file to read tag groups from. May be specified more than once (groups will be combined). Files may be in either the TagSpaces or the autotagical format.
• [-s/--schema <schema file>] -- Path to a schema file to move/rename files based on. May be specified more than once, in which case rules are prioritized in the order files are specified.

Functionality Options

These options tweak the default functioning of autotagical as specified below, typically adjusting how various circumstances are dealt with.

• [-A/--allmatchroot] -- Makes the root of an output folder match all tags, i.e. every single file will be moved to the output folder, even if it does not match anywhere more specifically. Use of this option is bad practice (consider using the /*| operator as a root filter instead), but it is provided for the user's convenience. This option does not imply -M, i.e. files that could not be renamed will not be moved to the root folder just because -A is set.
• [--cleanin] -- Clean up (delete) all empty folders in the input folder/s. This will recurse, whether or not the -R flag is set.
• [--cleanout] -- Clean up (delete) all empty folders in the output folder. This will recurse, whether or not the -R flag is set.
• [-c/--clean] -- Clean up (delete) all empty folders in both input and output folders.
• [-F/--failforcerename] -- This option flags failing to rename a manually-named file that is being forcibly renamed due to the -N option should be considered a failure to name the file. That sounds complicated, but consider these cases:
  • [-F] -- Normal behavior (option will be ignored).
  • [-N] -- Force rename manually-named files, but manual names are "good enough" and manually-named files that cannot be renamed will be moved.
  • [-F -N] -- Force rename manually-named files and treat failures as failures. Manually-named files that cannot be renamed will not be moved.
  • [-N -M] -- Force rename manually-named files. All files will be moved.
  • [-F -N -M] -- Equivalent to -N -M. -F has no effect.
• [-k/--keep] -- Keep original files in the input folders untouched, i.e copy files to their new destinations rather than move them.
• [-m/--move] -- Only move files into a directory structure, do not try to rename them.
• [-M/--moveall] -- Move all files, not only ones that are manually-named/ successfully renamed.
• [-n/--name] -- Only rename files, do not try to move them into any directory structure. All files will be placed in the root of the output folder.
• [-N/--renamemanual] -- Forcibly try to rename manually-named files, not just unnamed ones.
• [-t/--trial] -- Trial run. Do not actually move or rename files, just log what would happen. Combine with -v to check output before live run. Using this is good practice, especially after making any changes to a schema or options. The -t option ensures no changes will be reflected to disk whatsoever.

Logging Options

These options tweak what sorts of messages are displayed when autotagical is run (and whether to save them or just print them to the console).

• [--debug] -- Display absolutely everything. You should probably never use this.
• [-l/--log <log file>] -- Output messages to the specified file rather than just the console. By default, messages will be appended to the end of the file.
• [-L/--overwritelog] -- Overwrite the specified log file, rather than append to it. Has no effect without -l.
• [-P/--posix] -- Silence warnings specific to Windows. Use this only if the files are never to be used with Windows systems (which are pickier about what file names can contain).
• [-q/--quiet] -- Silence all warnings and only display actual errors. Use of this is not recommended, as warnings are typically printed for good reason.
• [-v/--verbose] -- Print all actions taken. This will list every file movement/renaming, rather than merely warning about failures. Most useful when combined with -t to check that a schema is doing what one wants before running it for real.

Unsafe Options

Do not use these unless you have very good reason to. Data loss can occur.

• [--force] -- Forcibly move/rename files, even if there is a file or directory in the way. This will clobber files; use at your own risk, as data loss can occur.
• [--yes] -- Assume "yes" for all user prompts. This implies --force and will clobber files and directories. Use at your own risk, as data loss can occur.

Setting Priority

autotagical will always preferentially use settings specified in the following order, from highest to lowest, with each overriding any settings from lower priority sources. Note that only one config file will be loaded (the first by priority).

1. Command-line arguments.
2. Config file (first found from below).
3. Config file loaded using -C/--config argument.
4. .autotagrc file in output folder (first folder containing one is used).
5. .autotagrc file in input folder (first folder containing one is used).
6. .autotagrc file in autotagical module folder.

Config File Format

.autotagrc (or any config file specified via command-line) should have the same format as any command line arguments one would otherwise care to pass. Unsafe options will be ignored in config files, i.e. --force and --yes will have no effect if set in a config file. This is to prevent data loss without explicit user input. Whitespace and newlines are ignored, e.g. one might write:

-H

  -P

in a config file to process hidden and ignore Windows-specific warnings.

Tag Group Format

autotagical is capable of reading the JSON files produced by exporting tag groups from TagSpaces. Alternately, tag groups may be defined in a more simple, somewhat more human-readable fashion in JSON. This autotagical tag group format supports additional features not available in the TagSpaces format, detailed below.

{
  "file_type": "autotagical_tag_groups",
  "tag_group_file_version": "1.1",
  "tag_groups": [
    {
      "name": "tag group 1",
      "tags": ["tag1", "tag2"...]
    },
    {
      "name": "tag group 2",
      "tags": ["tag3", "/G|tag group 1", "/RE|regex 1",...]
    },
  ]
}

Inheritance

In the autotagical format, tag groups support simple inheritance, where a child tag group may be defined in terms of one or more parent tag groups (as well as any additional tags). Inheritance is indicated by prefixing a group name with /G| in the tags property of a group. In such cases, the child group will inherit all tags in any of the parent groups. This is useful for simplicity; it ensures that one only has to manually define the most "leaf" tag groups while still writing filters based on more broad groups, all of which will update if the more refined groups are updated. Consider the following simple use case:

{
    "file_type": "autotagical_tag_groups",
    "tag_group_file_version": "1.1",
    "tag_groups": [
        {
            "name": "American Styles",
            "tags": [
                "ipa",
                "dipa",
                "pale_ale"
            ]
        },
        {
            "name": "Belgian Styles",
            "tags": [
                "witbier",
                "dubbel",
                "tripel"
            ]
        },
        {
            "name": "Beer",
            "tags": [
                "/G|American Styles",
                "/G|Belgian Styles"
            ]
        }
    ]
}

Here, we have defined the Beer group in terms of American Styles and Belgian Styles, rather than having to specify all 6 tags that should fall under it. Not only is this quicker to write, but it prevents errors of oversight when one decides to add tags later. If, for example, quadrupel is added to the Belgian Styles group, it will be part of the Beer group without one having to remember to manually add it in two places. While this simple case doesn't seem that unmanageable without inheritance, the number of places where duplicate tags have to be added can quickly spiral out of control with multiple levels of tag group.

Keep in mind the following details about inheritance:

• Multiple inheritance -- Tag groups can inherit from multiple parent group (as shown in the example).
• Multilevel inheritance -- Tag groups support multilevel inheritance, i.e. an Alcohol group might inherit from Beer which inherits from American Styles.
• The (lack of) diamond problem -- As there is no overriding in tag group inheritance, the diamond problem (where a grandparent is inherited via two different routes) is handled without problem.
• Circular inheritance -- It is okay for two groups to inherit from each other (whether via intermediaries or not). Each inheritance path will only be followed once, i.e. a tag group will stop "following" an inheritance path if it is instructed to inherit from itself.
• Flexible ordering -- Tag group inheritance is flexible in the order that groups are defined. There is no need for a tag group to be located after a group it inherits from. In fact, tag groups can inherit from groups in completely different tag group files, regardless of the order they are loaded in (so long as they are both loaded).

Regex Tag Groups

The autotagical tag group format supports defining tag groups in terms of regexes. This can be incredibly powerful in certain situations. For example, consider tagged bank statements. Rather than having to manually add every account number tag to an "Account Number" group, one can instead define a tag group as follows:

{
    "name": "Account Number",
    "tags": ["/RE|(?:xx|\\*\\*)[0-9]{4}"]
}

This will make any tag of the form **1234 or xx1234 match the "Account Number" tag group. If one deals with a large number of tags with a standardized form, this can be extremely valuable, especially if they change frequently.

Keep in mind the following details about regex tag groups:

• Full match -- Regexes are used with re.fullmatch() and as such must match 100% of the tag name.
• Inheritance -- Regexes will be inherited as with any other tags in a group.
• Mixed and matched -- A tag group can be defined in terms of any combination of inheritance, normal tags, or regexes. They do not need to be specified solely in terms of regexes.
• Tag in group (/?TIG|) operator -- Regex tag groups work normally with the tag in group format string operator. The first tag to match the regex will be returned for it.

Schema Format

A schema is defined in a human-readable fashion in JSON and should consist of a single object as follows:

{
  "file_type": "autotagical_schema",
  "schema_file_version": "1.1",
  "tag_formats": [],
  "unnamed_patterns": [],
  "renaming_schemas": [],
  "movement_schema": []
}

Each of these four keys should be assigned to an array, the structure of which is described in the following sections.

tag_formats

"tag_formats": [
  {
    "tag_pattern": "Regex containing groups: file, raw_tags, tags, and extension.",
    "tag_split_pattern": "Regex to split tags with"
  },
  ...
]

The various patterns must be valid Python regexes. "tag_pattern" will be used with re.fullmatch() and should match the entirety of a file name and contain the following named groups:

• file -- Matches the original file name (without tags and extension).
• raw_tags -- Matches the entirety of tag data to be preserved, including any delimeters/demarcating characters.
• tags -- Matches only the tags themselves and any separating characters.
• extension -- Matches the file extension (if any).

"tag_split_pattern" should match whatever delimiter separates individual tags and will be used with re.split(). An example is provided below, which matches the tag format used by TagSpaces:

"tag_formats": [
  {
    "tag_pattern": "(?P<file>.+)(?P<raw_tags>\\[(?P<tags>.+?)\\])(?P<extension>.*?)",
    "tag_split_pattern": "\\s+"
  }
]

More than one such set of patterns may be provided in a single tag_formats array, allowing autotagical to deal with files tagged in multiple formats in one run.

unnamed_patterns

"unnamed_patterns": [
  "regex pattern 1",
  "regex pattern 2",
  ...
]

The various patterns must be valid Python regexes. Each will be used with re.match() and should match the file names (less tags) of such files to be treated as requiring renaming. They will be matched against the concatenation of the file and extension groups produced by the use of "tag_pattern" above, so do not attempt to match against anything not captured by those two groups. An example is provided below, which might match PDF files with timestamps produced by two different scanners:

"unnamed_patterns": [
  "[0-9]{4}_[0-9]{2}_[0-9]{2}_[0-9]{2}_[0-9]{2}_[0-9]{2}\\s*.pdf",
  "(Pages\\sfrom)?\\s*XScan_[0-9]{14}\\s*.pdf"
]

renaming_schema

"renaming_schemas": [
  {
    "filter": ["condition 1", "condition 2", ...],
    "format_string": "file name format string"
  },
  ...
]

"renaming_schemas" is simply a list of filters and an explanation of how to name files matching any of them. This list is ordered, and files will be renamed according to the first they match, e.g. if a file has tag1 and tag2 and the first filter in the array matches tag2 and the second tag1, the file will be renamed according to tag2. This allows one to define priorities of renaming. See the Filters section for information on filters, conditions, and the various operators one may include in them.

"format_string" defines how to rename files matching the filter. It cannot contain the / character except where it denotes operators (as file names cannot contain /). At its simplest, it is simply a string that the file will be renamed to, but that string may include any number/combination of operators. See the Format Strings section for information on format strings and the various operators one may include in them.

movement_schema

"movement_schema": [
  {
    "filter": ["condition 1", "condition 2", ...],
    "subfolder": "<format string 1>",
    "sublevels": [
      {
        "filter": ["condition 3", "condition 4", ...],
        "subfolder": "<format string 2>",
        "sublevels": [...]
      },
      ...
    ]
  },
  ...
]

"movement_schema" defines an output folder hierarchy iteratively by nesting filters. At each folder level, multiple filters can exist that either pass files to lower subfolders or place them at the current level. These lists are ordered, and files will be sorted according to the first filter they match, e.g. if a file has tag1 and tag2 and at a level the first filter in the array matches tag2 and the second tag1, the file will be sorted according to tag2. This allows one to define priorities of sorting. See the Filters section for information on filters, conditions, and the various operators one may include in them.

If "subfolder" contains a format string, it will be interpreted and added to the folder hierarchy that the file will be placed in (and further sorted based on "sublevels"); if it is left blank "", files will be placed in the current (in the hierarchy) directory without further sorting. If "sublevels" is left empty [], files will be placed in the specified subfolder without further sorting. Note that a movement schema does not have to have a path for every possible file. Files that fail to "find a home" will be left in the input folder and a warning will be printed (unless -A is specified). See the Format Strings section for information on format strings and the various operators one may include in them.

Additionally, note that complete hierarchies (i.e. those that terminate with explicitly placing the file in a folder) will be preferred over partial hierarchies (i.e. if a file percolates some distance down a hierarchy and then matches no filters). In the event that no complete hierarchy can be found, the first partial one will be used to move the file. It is bad practice to rely on this behavior though; one should use the /*| operator if one explicitly wishes absolutely any file that reaches a filter level to reside there.

Filters

A filter (wherever it might show up in a schema) is defined by an array of condition sets. These condition sets are combined in the logical sense by inclusive or, i.e. matching at least one condition set is necessary and sufficient to match the overall filter. At their simplest, a condition set may simply be a tag, e.g. "filter": ["tag1", "tag2"] will match any file with either tag1 or tag2 (or both) on it. However, the following operators may be used to construct more complex condition sets (whether in filters or in the conditional /?| operator):

• /G| -- The prefix /G| is used to denote a tag group instead of a tag name, e.g. "filter": ["/G|Group 1", "tag2"] will match any file with at least one tag in Group 1 or the tag tag2 (or both).
• /*| -- The all operator /*| matches all files, regardless of how they are tagged.
• /&| -- /&| is a logical "and" operator, requiring matching both conditions, e.g. "filter": ["tag1/&|tag2"] will match files that have both tag1 and tag2. A condition may contain any number of /&| operators, e.g. one may create the condition "tag1/&|tag2/&|/G|Group 1".
• /!| -- The not prefix /!| negates the next condition. This prefix must come before any others logically, i.e. you must write /!|/G|<group> rather than /G|/!|<group> or /!|/*| rather than /*|/!|. The /!| operator can follow the logical "and" operator /&|, e.g. "<tag1>/&|/!|<tag2>", which will match any file that has tag1 and does not have tag2.

There are no (realistic) limits on the degree to which these operators may be combined or how many condition sets a filter might have.

Format Strings

A format string is simply a string that may or may not contain various operators. These operators will be replaced with the corresponding data when the format string is interpreted.

• /EXT| -- Anywhere it is put in the format string, /EXT| will be replaced with the original extension of the file, as defined by the extension group in the "tag_formats" regex that matched the file. This is obviously useful if you're renaming multiple types of file and want to preserve extensions. You will almost always want to end your format string with /EXT|.
• /FILE| -- Anywhere it is put in the format string, /FILE| will be replaced with the original name of the file, as defined by the file group in the "tag_formats" regex that matched the file.
• /TAGS| -- Anywhere it is put in the format string, /TAGS| will be replaced with the tags on the original file. This is necessary to avoid your renamed files becoming de-tagged and should almost always be included in a renaming format string.
• /?|<condition>/T|<true text>/F|<false text>/E?| -- The conditional operator /?| allows for conditional naming. If <filter> is matched, the entire expression will be replaced with <true text>; if it does not match, the entire expression will be replaced with <false text>. Either text (or both, but why would you) may be empty. The conditional operator can take anything that can be in a filter condition. See the Filters section for information on filters, conditions, and the various operators one may include in them. <true text> and <false text> may contain other operators, i.e. /EXT|, /FILE|, /TAGS|, and /ITER|. Note: conditional operators cannot be nested or contain /?T|<tag>/| or /?G|<tag group>/| within replacement text.
• /?T|<tag>/| -- The tag conditional operator /?T| will insert the literal name of the tag <tag> if it is present on the file. Note that this is equivalent to /?|<tag>/T|/<tag>/F|/E?|; it is merely a shortcut.
• /?G|<tag group>/| -- The tag group conditional operator /?G| will insert the literal name of the tag group <tag group> if one of its tags is present on the file. Note that this is equivalent to /?|/G|<tag group>/T|<tag group>/F|/E?|; it is merely a shortcut.
• /?TIG|<tag group>/| -- The "tag in group" operator is a special operator that will insert a tag on the file that is in the specified tag group. For example, /?TIG|group1?/| will resolve to tag1 if tag1 is on the file and in group1, or tag2 if tag2 is on the file and in group1. If multiple tags on the file are in the group, the first (in tagging order, left-to-right) will be substituted in. If the file lacks a tag in the group, the entire operator will simply be blank. This "tag in group" operator is most useful for writing flexible schema where the exact tags cannot be predicted in advance. For example, if one wanted to sort bank statements into folders based on what account they are associated with, one could use the "tag in group" operator with an "Account Number" group and then only have to update the tag group with the various account number tags, rather than having to specify every account number in the schema. As an example, consider the following cases, with Group 1 = tag1, tag3 and format_string = "Tag: /?TIG|Group 1/|":
  • File [tag1, tag2] -- "Tag: tag1"
  • File [tag2, tag4] -- "Tag: "
  • File [tag1, tag2, tag3] -- "Tag: tag2"
• /ITER|<text>/#|<other text>/EITER| -- The /ITER| operator is complicated but important. It is invoked only in the event that multiple files are going to be renamed to the same name. In this case, the text is placed in the file name, along with /#| being replaced by the n-th file that this is that has had the same name. Otherwise, the entirety of the /ITER| operator is ignored. The /ITER| operator should not be used in folder name format strings. It will be ignored. In essence, the /ITER| tag "counts" how many times the same file name has been produced. It is good practice to always include an /ITER| operator in your schema to avoid files not being renamed due to potential clobbering. /#| may appear more than once in an /ITER| operator, but there is usually no need to. The /ITER| operator may contain any other operator, including the conditional operator /?|, but may not be nested. Note that the /ITER| operator will not be used if files end up with the same name but different output directories. It will only appear if necessary to avoid clobbering. An example will make this easier to understand. Consider the format string Widget/ITER| /#|/EITER| in the following cases:
  • 1 matching file -- The file will be named Widget.
  • 3 matching files in same folder -- The files will be named Widget 1, Widget 2, and Widget 3.
  • 3 matching files, each in a different folder -- The files will all be named Widget.

Known Issues

• Only POSIX hidden files are considered hidden, i.e. those that begin with a . dot, not those hidden as Windows does it. This is most likely a Won’t Fix.

Tests

autotagical may be tested by cloning this repository and running:

python setup.py test

from within the root directory. Note that this may require python3 instead of python, depending on your python installation.

Authors

• SiriusStarr

License

This project is licensed under the GNU General Public License v3.0 - see the LICENSE.md file for details