File size: 6,161 Bytes
1e3b872 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 |
__author__ = "receyuki"
__filename__ = "invokeai.py"
__copyright__ = "Copyright 2023"
__email__ = "receyuki@gmail.com"
import json
import re
from ..format.base_format import BaseFormat
from ..utility import remove_quotes
class InvokeAI(BaseFormat):
SETTING_KEY_INVOKEAI_METADATA = [
"",
("scheduler", "refiner_scheduler"),
"seed",
("cfg_scale", "refiner_cfg_scale"),
("steps", "refiner_steps"),
"",
]
SETTING_KEY_METADATA = [
"model_weights",
"sampler",
"seed",
"cfg_scale",
"steps",
"",
]
SETTING_KEY_DREAM = ["", "A", "S", "C", "s", ""]
DREAM_MAPPING = {
"Steps": "s",
"Seed": "S",
"Size": "",
"CFG scale": "C",
"Sampler": "A",
}
def __init__(self, info: dict = None, raw: str = ""):
super().__init__(info, raw)
if "invokeai_metadata" in self._info:
self._invoke_invoke_metadata()
elif "sd-metadata" in self._info:
self._invoke_metadata()
elif "Dream" in self._info:
self._invoke_dream()
def _invoke_invoke_metadata(self):
data_json = json.loads(self._info.get("invokeai_metadata"))
self._positive = data_json.pop("positive_prompt").strip()
self._negative = data_json.pop("negative_prompt").strip()
self._raw = "\n".join([self._positive, self._negative, str(data_json)])
self._setting = remove_quotes(str(data_json)).strip("{ }")
self._width = str(data_json.get("width"))
self._height = str(data_json.get("height"))
has_refiner = True if data_json.get("refiner_model") else False
for p, s in zip(super().PARAMETER_KEY, InvokeAI.SETTING_KEY_INVOKEAI_METADATA):
match p:
case "model":
self._parameter[p] = remove_quotes(
str(
(
data_json.get("model").get("model_name"),
data_json.get("refiner_model").get("model_name"),
)
)
if has_refiner
else str(data_json.get("model").get("model_name"))
)
case "seed":
self._parameter["seed"] = str(data_json.get("seed"))
case "size":
self._parameter["size"] = (
str(data_json.get("width")) + "x" + str(data_json.get("height"))
)
case _:
self._parameter[p] = remove_quotes(
str((data_json.get(s[0]), data_json.get(s[1])))
if has_refiner
else str(data_json.get(s[0]))
)
def _invoke_metadata(self):
data_json = json.loads(self._info.get("sd-metadata"))
image = data_json.pop("image")
prompt = (
image.get("prompt")[0].get("prompt")
if isinstance(image.get("prompt"), list)
else image.get("prompt")
)
self._positive, self._negative = self.split_prompt(prompt)
raw_list = [
item
for item in [
self._positive,
self._negative,
self._info.get("Dream"),
self._info.get("sd-metadata"),
]
if item != ""
]
self._raw = "\n".join(raw_list).strip()
image.pop("prompt")
self._setting = remove_quotes(
", ".join([str(data_json).strip("{ }"), str(image).strip("{ }")])
)
self._width = str(image.get("width"))
self._height = str(image.get("height"))
for p, s in zip(super().PARAMETER_KEY, InvokeAI.SETTING_KEY_METADATA):
match p:
case "model":
self._parameter["model"] = data_json.get(s)
case "size":
self._parameter["size"] = (
str(image.get("width")) + "x" + str(image.get("height"))
)
case _:
self._parameter[p] = str(image.get(s))
def _invoke_dream(self):
data = self._info.get("Dream")
# match parameters like '"Prompt"Setting'
pattern = r'"(.*?)"\s*(.*?)$'
prompt, setting = re.search(pattern, data).groups()
self._positive, self._negative = self.split_prompt(prompt.strip('" '))
self._raw = "\n".join([self._positive, self.negative, self._info.get("Dream")])
# match parameters like "-s 30"
pattern = r"-(\w+)\s+([\w.-]+)"
setting_dict = dict(re.findall(pattern, setting))
setting_list = []
for key, value in InvokeAI.DREAM_MAPPING.items():
if key == "Size":
setting_list.append(
key + ": " + setting_dict.get("W") + "x" + setting_dict.get("H")
)
else:
setting_list.append(key + ": " + setting_dict.get(value))
self._setting = ", ".join(setting_list)
self._width = str(setting_dict.get("W"))
self._height = str(setting_dict.get("H"))
for p, s in zip(super().PARAMETER_KEY, InvokeAI.SETTING_KEY_DREAM):
match p:
case "model":
self._parameter["model"] = ""
case "size":
self._parameter["size"] = (
str(setting_dict.get("W")) + "x" + str(setting_dict.get("H"))
)
case _:
self._parameter[p] = setting_dict.get(s)
@staticmethod
def split_prompt(prompt: str):
# match parameters like "Positive[Negative]"
pattern = r"^(.*?)\[(.*?)\]$"
match = re.match(pattern, prompt)
if match:
positive, negative = match.groups()
positive = positive.strip()
negative = negative.strip()
else:
positive = prompt.strip()
negative = ""
return positive, negative
|