Add: dev_tools item statistics and doc

- add basic item templates

dev_tools/item_stastistics.py

@@ -0,0 +1,241 @@
+import os
+
+# os.chdir('../')
+import module.config.server as server
+
+server.server = 'cn'  # Don't need to edit, it's used to avoid error.
+print(os.getcwd())
+
+import numpy as np
+from PIL import Image
+import cv2
+import time
+from module.combat.assets import GET_ITEMS_1, GET_ITEMS_2
+from module.handler.assets import INFO_BAR_1
+from module.base.button import ButtonGrid, Button
+from module.base.ocr import Ocr
+from module.logger import logger
+
+
+"""
+Set your folder here
+Examples: xxx/campaign_7_2
+"""
+IMAGE_FOLDER = ''
+STATUS_ITEMS_INTERVAL = 10
+TEMPLATE_THRESHOLD = 0.9
+
+BATTLE_STATUS_FOLDER = f'{IMAGE_FOLDER}/status'
+GET_ITEMS_FOLDER = f'{IMAGE_FOLDER}/get_items'
+TEMPLATE_FOLDER = f'{IMAGE_FOLDER}/item_template'
+for f_ in [TEMPLATE_FOLDER]:
+    if not os.path.exists(f_):
+        os.mkdir(f_)
+BATTLE_STATUS_TIMESTAMP = np.array([int(f.split('.')[0]) for f in os.listdir(BATTLE_STATUS_FOLDER)])
+ITEM_GRIDS_1_ODD = ButtonGrid(origin=(336, 298), delta=(128, 0), button_shape=(96, 96), grid_shape=(5, 1))
+ITEM_GRIDS_1_EVEN = ButtonGrid(origin=(400, 298), delta=(128, 0), button_shape=(96, 96), grid_shape=(4, 1))
+ITEM_GRIDS_2 = ButtonGrid(origin=(336, 227), delta=(128, 142), button_shape=(96, 96), grid_shape=(5, 2))
+ENEMY_GENRE_BUTTON = Button(area=(782, 285, 961, 319), color=(), button=(), name='ENEMY_GENRE')
+
+
+class AmountOcr(Ocr):
+    def ocr(self, image):
+        start_time = time.time()
+
+        image_list = [self.pre_process(i) for i in image]
+        result_list = self.cnocr.ocr_for_single_lines(image_list)
+        result_list = [self.after_process(result) for result in result_list]
+
+        if len(self.buttons) == 1:
+            result_list = result_list[0]
+        logger.attr(name='%s %ss' % (self.name, str(round(time.time() - start_time, 3)).ljust(5, '0')),
+                    text=str(result_list))
+
+        return result_list
+
+    def after_process(self, raw):
+        """
+        Returns:
+            int:
+        """
+        raw = super().after_process(raw)
+        if not raw:
+            result = 0
+        else:
+            result = int(raw)
+
+        return result
+
+
+AMOUNT_OCR = AmountOcr([], back=(-200, -200, -200), lang='digit', name='Amount_ocr')
+ENEMY_GENRE_OCR = Ocr(ENEMY_GENRE_BUTTON, lang='cnocr', use_binary=False, back=(127, 127, 127))
+
+
+class ImageError(Exception):
+    pass
+
+
+class ItemTemplate:
+    def __init__(self, image):
+        self.image = np.array(image)
+
+    def match(self, image):
+        res = cv2.matchTemplate(self.image, np.array(image), cv2.TM_CCOEFF_NORMED)
+        _, similarity, _, _ = cv2.minMaxLoc(res)
+        return similarity > TEMPLATE_THRESHOLD
+
+    def save(self, name):
+        image = Image.fromarray(self.image)
+        image.save(f'{TEMPLATE_FOLDER}/{name}.png')
+
+
+class ItemTemplateGroup:
+    def __init__(self):
+        self.templates = {}
+        for file in os.listdir(TEMPLATE_FOLDER):
+            name = file[:-4]
+            image = Image.open(f'{TEMPLATE_FOLDER}/{file}').convert('RGB')
+            self.templates[name] = ItemTemplate(image)
+
+    def match(self, item):
+        for name, template in self.templates.items():
+            if template.match(item.image):
+                return name
+
+        template = ItemTemplate(item.get_template())
+        name = [int(n) for n in self.templates.keys() if n.isdigit()]
+        if len(name):
+            name = str(max(name) + 1)
+        else:
+            name = str(len(self.templates.keys()) + 1)
+
+        logger.info(f'New item template: {name}')
+        self.templates[name] = template
+        template.save(name)
+        return name
+
+
+template_group = ItemTemplateGroup()
+
+
+class Item:
+    def __init__(self, image):
+        self.image = image
+        self.is_valid = np.mean(np.array(image.convert('L')) > 127) > 0.1
+        self.name = 'Default_item'
+        self.amount = 1
+        if self.is_valid:
+            self.name = template_group.match(self)
+            if not self.name.startswith('_') and '_' in self.name:
+                self.name = '_'.join(self.name.split('_')[:-1])
+
+    def __str__(self):
+        return f'{self.name}_x{self.amount}'
+
+    @property
+    def has_amount(self):
+        return 'T' in self.name or self.name == '物资'
+
+    def get_template(self):
+        # return self.image.crop((5, 5, 90, 68))
+        return self.image.crop((40, 21, 89, 70))
+
+    def get_amount(self):
+        return self.image.crop((60, 75, 91, 88))
+
+
+class Items:
+    def __init__(self, timestamp):
+        self.timestamp = timestamp
+        self.get_items = Image.open(f'{GET_ITEMS_FOLDER}/{timestamp}.png').convert('RGB')
+
+        # Enemy genre
+        interval = np.abs(BATTLE_STATUS_TIMESTAMP - timestamp)
+        if np.min(interval) > STATUS_ITEMS_INTERVAL * 1000:
+            raise ImageError(f'Timestamp: {timestamp}, battle_status image not found.')
+        self.status_timestamp = BATTLE_STATUS_TIMESTAMP[np.argmin(interval)]
+        self.enemy = 'Default_enemy'
+
+        # get_item image properties
+        if INFO_BAR_1.appear_on(self.get_items):
+            raise ImageError(f'Timestamp: {timestamp}, Info bar')
+        if GET_ITEMS_1.appear_on(self.get_items):
+            self.row = 1
+            self.is_odd = self.get_is_odd(self.get_items)
+            self.grids = ITEM_GRIDS_1_ODD if self.is_odd else ITEM_GRIDS_1_EVEN
+        elif GET_ITEMS_2.appear_on(self.get_items):
+            self.row = 2
+            self.is_odd = True
+            self.grids = ITEM_GRIDS_2
+        else:
+            raise ImageError(f'Timestamp: {timestamp}, Image is not a get_items image.')
+
+        # Crop items
+        self.items = []
+        for button in self.grids.buttons():
+            item = Item(self.get_items.crop(button.area))
+            if item.is_valid:
+                self.items.append(item)
+
+    @staticmethod
+    def get_is_odd(image):
+        image = image.crop((628, 294, 651, 396))
+        return np.mean(np.array(image.convert('L')) > 127) > 0.1
+
+    def predict(self):
+        self.battle_status = Image.open(f'{BATTLE_STATUS_FOLDER}/{self.status_timestamp}.png').convert('RGB')
+        self.enemy = ENEMY_GENRE_OCR.ocr(self.battle_status)
+        enemy = self.enemy
+        # Delete wrong OCR result
+        for letter in '-一个―~(':
+            enemy = enemy.replace(letter, '')
+        self.enemy = enemy
+
+        amount_items = [item for item in self.items if item.has_amount]
+        amount = AMOUNT_OCR.ocr([item.get_amount() for item in amount_items])
+        for a, i in zip(amount, amount_items):
+            i.amount = a
+
+    def get_data(self):
+        return [[self.timestamp, self.status_timestamp, self.enemy, item.name, item.amount] for item in self.items]
+
+
+"""
+These code is for testing
+Set your image name here
+Examples: 159022xxxxxxx (int)
+"""
+# ts = 1590227624035
+# items = Items(ts)
+# for item in items.items:
+#     print(item.amount, item.name)
+
+"""
+These code is for template extracting
+"""
+# from tqdm import tqdm
+# for ts in tqdm([int(f.split('.')[0]) for f in os.listdir(GET_ITEMS_FOLDER)]):
+#     try:
+#         items = Items(ts)
+#     except Exception:
+#         logger.warning(f'Error image: {ts}')
+#         continue
+
+"""
+These code is for final statistic
+Set your csv file name here
+Examples: c64.csv
+"""
+# csv_file = 'c64.csv'
+# import csv
+# from tqdm import tqdm
+# with open(csv_file, 'a', newline='') as file:
+#     writer = csv.writer(file)
+#     for ts in tqdm([int(f.split('.')[0]) for f in os.listdir(GET_ITEMS_FOLDER)]):
+#         try:
+#             items = Items(ts)
+#             items.predict()
+#             writer.writerows(items.get_data())
+#         except Exception:
+#             logger.warning(f'Error image: {ts}')
+#             continue

doc/item_statistics_en.md

@@ -0,0 +1,141 @@
+# How to do statistics on item drop rate
+
+This document will show how to use `dev_tools/item_statistics.py`
+
+## Enable statistics in alas
+
+In alas GUI, 
+
+- set `enable_drop_screenshot` to `yes`, if enabled, alas will add a 1s sleep before screenshot, to avoid capture the flash. There's a flash light when item shows up.
+- set `drop_screenshot_folder` to be the folder you want to save. It is recommended to save it in SSD.
+
+After few hours or few days of running, you will get a folder structure like:
+
+```
+<your_folder>
+    campaign_7_2
+        get_items
+            158323xxxxxxx.png
+            158323xxxxxxx.png
+            158323xxxxxxx.png
+        get_mission
+        get_ship
+        mystery
+        status
+    campaign_10_4_HARD
+        get_items
+        get_mission
+        get_ship
+        status
+    d3
+        get_items
+        get_mission
+        get_ship
+        status
+```
+
+Screenshot are named after millesecond timestamp.
+
+## Prepare a new environment
+
+- Prepare another virtual environment, accoring to `requirements.txt`. But use the GPU version of `mxnet`.
+
+   I am using GTX1080Ti, and I installed `mxnet-cu80==1.4.1`, `CUDA8.0`, `cuDNN`. Google  `mxnet gpu install`, and see how to do in details. You may intall other version of CUDA, and mxnet for that CUDA, because you are using another graphic card.
+
+- Look for the cnocr in your virtual environment. Replace site-packages\cnocr\cn_ocr.py line 89
+
+```
+mod = mx.mod.Module(symbol=sym, context=mx.cpu(), data_names=data_names, label_names=None)
+```
+
+​	to be:
+
+```
+mod = mx.mod.Module(symbol=sym, context=mx.gpu(), data_names=data_names, label_names=None)
+```
+
+​	Now cnocr will run on GPU.
+
+​	You can skip these anyway, and use the same environment as alas, but the OCR will run really slow.
+
+- Install tqdm, a package to show progressbar.
+
+```
+pip install tqdm
+```
+
+## Extract item_template
+
+Copy folder `dev_tools\item_template` to the map folder such as `<your_folder>\campaign_7_2`.
+
+Change the folder in line 24
+
+These template are named in chinese, rename them in English.
+
+>**How to a name template image**
+>
+>You should use their full name, such as "138.6mm单装炮Mle1929T3", instead of short name or nickname, such as "DD_gun".
+>
+>If you have same item with different image, use names like `torpedo_part.png`, `torpedo_part_2.png`, they will a classified as torpedo_part
+
+Uncomment the part for item extract in dev_tools/item_statistics.py, and run, you will have some new item templates. Here's an example log:
+
+```
+  1%|          | 107/12668 [00:05<10:24, 20.10it/s]2020-06-03 10:39:42.609 | INFO | New item template: 50
+  1%|          | 158/12668 [00:07<10:42, 19.47it/s]2020-06-03 10:39:45.098 | INFO | New item template: 51
+  2%|▏         | 207/12668 [00:10<10:33, 19.66it/s]2020-06-03 10:39:47.772 | INFO | New item template: 52
+  2%|▏         | 215/12668 [00:10<11:20, 18.29it/s]2020-06-03 10:39:48.304 | INFO | New item template: 53
+100%|██████████| 12668/12668 [10:33<00:00, 19.99it/s]
+```
+
+Rename those new templates.
+
+If you find some items haven't been extracted,  try use line 140, instead of 141.
+
+## Final statistic
+
+Uncomment the part for final statistic, configure the csv file you wang to save.
+
+The ocr model may not works fine in EN.
+
+Here's an example log:
+
+```
+2020-06-03 12:23:55.355 | INFO | [ENEMY_GENRE 0.007s] 中型侦查舰队
+2020-06-03 12:23:55.363 | INFO | [Amount_ocr 0.009s] [1, 1, 22]
+100%|█████████▉| 14916/14919 [20:32<00:00, 13.20it/s]2020-06-03 12:23:55.442 | INFO | [ENEMY_GENRE 0.007s] 大型航空舰队
+2020-06-03 12:23:55.455 | INFO | [Amount_ocr 0.013s] [1, 1, 1, 17]
+2020-06-03 12:23:55.539 | INFO | [ENEMY_GENRE 0.007s] 敌方旗舰
+2020-06-03 12:23:55.549 | INFO | [Amount_ocr 0.010s] [1, 2, 1, 63]
+100%|█████████▉| 14918/14919 [20:33<00:00, 12.35it/s]2020-06-03 12:23:55.623 | INFO | [ENEMY_GENRE 0.007s] 精英舰队
+2020-06-03 12:23:55.633 | INFO | [Amount_ocr 0.010s] [1, 1, 1, 17]
+100%|██████████| 14919/14919 [20:33<00:00, 12.10it/s]
+```
+
+Now you got a csv file, formated to be:
+
+```
+<get_item_timestamp>, <battle_status_timestamp>, <enemy_genre>, <item_name>, <item_amount>
+```
+
+like this:
+
+```
+1590271317900,1590271315841,中型主力舰队,主炮部件T3,1
+1590271317900,1590271315841,中型主力舰队,物资,23
+1590271359374,1590271357251,小型侦查舰队,通用部件T1,1
+1590271359374,1590271357251,小型侦查舰队,鱼雷部件T2,1
+1590271359374,1590271357251,小型侦查舰队,物资,13
+1590271415308,1590271413207,敌方旗舰,彗星,1
+1590271415308,1590271413207,敌方旗舰,通用部件T3,1
+1590271415308,1590271413207,敌方旗舰,科技箱T1,1
+1590271415308,1590271413207,敌方旗舰,物资,42
+1590271415308,1590271413207,敌方旗舰,_比萨研发物资,1
+1590271415308,1590271413207,敌方旗舰,_鸢尾之印,1
+```
+
+You can open it in Excel or load it into database.
+
+## Improvement
+
+These code is running on single thread, you can try adding multiprocess to speed up. I didn't do that because it's still acceptable (20it/s without ocr, 12it/s with ocr)