update inferface

face_similarity/diary_cover_similarity.py

@@ -7,10 +7,24 @@ import dlib
 import faiss
 import numpy as np
 from gm_rpcd.all import bind
+from utils.cache import redis_client3
+from utils.es import es_query
+from utils.images import face_to_vec, url_to_ndarray
 
-from face_similarity.utils.cache import redis_client3
-from face_similarity.utils.es import es_query
-from face_similarity.utils.images import face_to_vec, url_to_ndarray
+base_dir = os.getcwd()
+print("base_dir: " + base_dir)
+model_dir = os.path.join(base_dir, "_models")
+facerec_model_path = os.path.join(model_dir, "dlib_face_recognition_resnet_model_v1.dat")
+shape_model_path = os.path.join(model_dir, "shape_predictor_68_face_landmarks.dat")
+faiss_index_path = os.path.join(base_dir, "_index", "diary_cover.index")
+
+diary_after_cover_vec_file = "./diary_after_cover_vec.txt"
+
+face_rec = dlib.face_recognition_model_v1(facerec_model_path)
+face_detector = dlib.get_frontal_face_detector()
+shape_predictor = dlib.shape_predictor(shape_model_path)
+FACE_TO_VEC_FUN = lambda img: face_to_vec(img, face_rec, face_detector, shape_predictor)
+FAISS_DIARY_INDEX = faiss.read_index(faiss_index_path)
 
 
 @bind("hello")
@@ -18,7 +32,7 @@ def hello():
     return ["hello", "world"]
 
 
-def save_diary_image_info(save_file, face_to_vec_f):
+def save_diary_image_info(save_file):
     q = {
         "query": {
             "bool": {
@@ -72,7 +86,7 @@ def save_diary_image_info(save_file, face_to_vec_f):
             if img.any():
                 count += 1
                 print("count: " + str(count) + " " + str(diary_id))
-                faces = face_to_vec_f(img)
+                faces = FACE_TO_VEC_FUN(img)
                 for face in faces:
                     line = str(diary_id) + "\t" + face["feature"] + "\n"
                     # print(line)
@@ -103,13 +117,13 @@ def save_faiss_index(load_file, save_path):
         print("faiss index saved")
 
 
-def get_similar_diary_ids_by_url(url, index, face_to_vec_f, limit=0.1):
+def get_similar_diary_ids_by_url(url, limit=0.1):
     img = url_to_ndarray(url)
     if img.any():
-        faces = face_to_vec_f(img)
+        faces = FACE_TO_VEC_FUN(img)
         for face in faces:
             face_feature = np.array(json.loads(face["feature"])).astype("float32")
-            _scores, _ids = index.search(np.array([face_feature]), 10)
+            _scores, _ids = FAISS_DIARY_INDEX.search(np.array([face_feature]), 10)
             ids = _ids.flat
             scores = _scores.flat
             tmp = list(set(zip(ids, scores)))
@@ -123,9 +137,9 @@ def get_similar_diary_ids_by_url(url, index, face_to_vec_f, limit=0.1):
         return []
 
 
-def get_similar_diary_ids_by_face_features(feature, index, face_to_vec_f, limit=0.1):
+def get_similar_diary_ids_by_face_features(feature, limit=0.1):
     feature = np.array(feature).astype("float32")
-    _sources, _ids = index.search(np.array([feature]), 10)
+    _sources, _ids = FAISS_DIARY_INDEX.search(np.array([feature]), 10)
     ids = _ids.flat
     scores = _sources.flat
     tmp = list(set(zip(ids, scores)))
@@ -137,56 +151,40 @@ def get_similar_diary_ids_by_face_features(feature, index, face_to_vec_f, limit=
     return res
 
 
-def save_diary_similarity(load_file, index_path, face_to_vec_f):
-    res_dict = {}
-    with open(load_file) as f:
-        lines = f.readlines()
-        print("lines: " + str(len(lines)))
-        count = 0
-        for line in lines:
-            count += 1
-            tmp = line.split("\t")
-            id = tmp[0]
-            feature = np.array(json.loads(tmp[1]))
-            print("{} {}".format(count, id))
-            tup_res = get_similar_diary_ids_by_face_features(feature, index_path, face_to_vec_f)
-            if tup_res:
-                res_dict[id] = json.dumps(tup_res)
+# def save_diary_similarity(load_file, index_path):
+#     res_dict = {}
+#     with open(load_file) as f:
+#         lines = f.readlines()
+#         print("lines: " + str(len(lines)))
+#         count = 0
+#         for line in lines:
+#             count += 1
+#             tmp = line.split("\t")
+#             id = tmp[0]
+#             feature = np.array(json.loads(tmp[1]))
+#             print("{} {}".format(count, id))
+#             tup_res = get_similar_diary_ids_by_face_features(feature, index_path, FACE_TO_VEC_FUN)
+#             if tup_res:
+#                 res_dict[id] = json.dumps(tup_res)
 
-    print("done: " + str(len(res_dict)))
-    key = random.choice(list(res_dict.keys()))
-    print(key + str(res_dict[key]))
+#     print("done: " + str(len(res_dict)))
+#     key = random.choice(list(res_dict.keys()))
+#     print(key + str(res_dict[key]))
 
-    redis_key = "doris:diary:face_similary"
-    redis_client3.hmset(redis_key, res_dict)
+#     redis_key = "doris:diary:face_similary"
+#     redis_client3.hmset(redis_key, res_dict)
 
 
 def main():
-    base_dir = os.getcwd()
-    print("base_dir: " + base_dir)
-    model_dir = os.path.join(base_dir, "_models")
-    facerec_model_path = os.path.join(model_dir, "dlib_face_recognition_resnet_model_v1.dat")
-    shape_model_path = os.path.join(model_dir, "shape_predictor_68_face_landmarks.dat")
-    faiss_index_path = os.path.join(base_dir, "_index", "current.index")
-
-    diary_after_cover_vec_file = "./diary_after_cover_vec.txt"
-
-    face_rec = dlib.face_recognition_model_v1(facerec_model_path)
-    face_detector = dlib.get_frontal_face_detector()
-    shape_predictor = dlib.shape_predictor(shape_model_path)
-    face_to_vec_f = lambda img: face_to_vec(img, face_rec, face_detector, shape_predictor)
-
-    # save_diary_image_info(diary_after_cover_vec_file, face_to_vec_f)
+    # save_diary_image_info(diary_after_cover_vec_file)
     # save_faiss_index(diary_after_cover_vec_file, faiss_index_path)
 
-    faiss_index = faiss.read_index(faiss_index_path)
-
     imgs = [
         "https://pic.igengmei.com/2020/07/03/1437/1b9975bb0b81-w", "https://pic.igengmei.com/2020/07/01/1812/ca64827a83da-w",
         "https://pic.igengmei.com/2020/07/04/1711/24f4131a9b1e-w", "https://pic.igengmei.com/2020/07/04/1507/e17a995be219-w"
     ]
     for img_url in imgs:
-        res = get_similar_diary_ids_by_url(img_url, faiss_index, face_to_vec_f, limit=0.18232107)
+        res = get_similar_diary_ids_by_url(img_url, limit=0.18232107)
         print(res)
 
     print("@@@@@@@@")
@@ -218,11 +216,9 @@ def main():
         -0.015351934358477592, -0.04335442930459976, -0.26258283853530884, -0.021509556099772453, 0.12185295671224594,
         -0.011788002215325832, 0.01337978895753622, -0.008025042712688446
     ]
-    res = get_similar_diary_ids_by_face_features(a, faiss_index, face_to_vec_f)
+    res = get_similar_diary_ids_by_face_features(a)
     print(res)
 
-    # # save_diary_similarity(diary_after_cover_vec_file, faiss_index, face_to_vec_f)
-
 
 if __name__ == "__main__":
     begin_time = time.time()

utils/admin.py

+from django.contrib import admin
+
+# Register your models here.

utils/apps.py

+from django.apps import AppConfig
+
+
+class UtilsConfig(AppConfig):
+    name = 'utils'

utils/models.py

+from django.db import models
+
+# Create your models here.

utils/tests.py

+from django.test import TestCase
+
+# Create your tests here.

utils/views.py

+from django.shortcuts import render
+
+# Create your views here.