fix chrome tasks (#230)

* fix chrome * fix: fix proxy setup * feat&fix: add proxy support in setup and remove hardcoded proxy from example * fix tasks * fix chrome finished * fix * clean chrome_fix code * clean chrome_fix code --------- Co-authored-by: adlsdztony <zzl0712@connect.hku.hk>
2025-07-03 21:32:41 +08:00
parent bdaf37e0e5
commit b2fb8b4222
26 changed files with 444 additions and 140 deletions
--- a/PUBLIC_EVALUATION_GUIDELINE.md
+++ b/PUBLIC_EVALUATION_GUIDELINE.md
@@ -172,4 +172,4 @@ For more, see: [MONITOR_README](./monitor/README.md)

 ### 4.2 VNC Remote Desktop Access

-You can also access Client instances via VNC at`http://<client-public-ip>:5090/vnc.html`
+You can also access Client instances via VNC at`http://<client-public-ip>:5910/vnc.html`
--- a/desktop_env/evaluators/getters/init.py
+++ b/desktop_env/evaluators/getters/init.py
@@ -25,7 +25,8 @@ from .chrome import (
    get_url_dashPart,
    get_active_url_from_accessTree,
    get_find_installed_extension_name,
-    get_info_from_website
+    get_info_from_website,
+    get_url_path_parse
 )
 from .file import get_cloud_file, get_vm_file, get_cache_file, get_content_from_vm_file
 from .general import get_vm_command_line, get_vm_terminal_output, get_vm_command_error
--- a/desktop_env/evaluators/getters/chrome.py
+++ b/desktop_env/evaluators/getters/chrome.py
@@ -1181,7 +1181,7 @@ def get_active_tab_html_parse(env, config: Dict[str, Any]):
                ], "shell": False})

            headers = {"Content-Type": "application/json"}
-            requests.post("http://" + host + ":" + server_port + "/setup" + "/launch", headers=headers, data=payload)
+            requests.post("http://" + host + ":" + str(server_port) + "/setup" + "/launch", headers=headers, data=payload)
            time.sleep(5)
            browser = p.chromium.connect_over_cdp(remote_debugging_url)
        target_page = None
@@ -1204,6 +1204,32 @@ def get_active_tab_html_parse(env, config: Dict[str, Any]):
            elements = target_page.query_selector_all(selector)
            return [element.text_content().strip() for element in elements if element]

+        def safely_get_direct_text_nodes_playwright(selector):
+            """
+            Extract all direct text node contents under the specified selector element (excluding text inside child div, span, etc.).
+            Returns a list of lists, each sublist contains the direct text nodes of one element.
+            Suitable for structures like: <div>SEA<div class="aura-separator"></div>NYC</div>
+            """
+            elements = target_page.query_selector_all(selector)
+            results = []
+            for element in elements:
+                texts = element.evaluate('''
+                    (node) => Array.from(node.childNodes)
+                        .filter(n => n.nodeType === Node.TEXT_NODE)
+                        .map(n => n.textContent.trim())
+                        .filter(Boolean)
+                ''')
+                results.append(texts)
+            return results[0]
+        
+        def safely_get_direct_li_playwright(selector):
+            elements = target_page.query_selector_all(selector + " li.catAllProducts")
+            return [element.query_selector('span').inner_text().strip() for element in elements if element.query_selector('span')]
+
+        def safely_get_only_child_text_content(selector):
+            elements = target_page.query_selector_all(selector)
+            return [element.query_selector('h3').text_content().strip() for element in elements if element.query_selector('h3')]
+
        if config["category"] == "class":
            class_multiObject = config.get("class_multiObject", {})
            for class_name, object_dict in class_multiObject.items():
@@ -1212,6 +1238,41 @@ def get_active_tab_html_parse(env, config: Dict[str, Any]):
                    index = int(order_key)
                    if len(elements_texts) > index:
                        return_json[key] = elements_texts[index]
+                        
+            class_multiObject_child = config.get("class_multiObject_child", {})
+            for class_name, object_dict in class_multiObject_child.items():
+                elements_texts = safely_get_direct_text_nodes_playwright("." + class_name)
+                for order_key, key in object_dict.items():
+                    index = int(order_key)
+                    if len(elements_texts) > index:
+                        return_json[key] = elements_texts[index]
+            
+            class_multiObject_only_child = config.get("class_multiObject_only_child", {})
+            for class_name, object_dict in class_multiObject_only_child.items():
+                elements_texts = safely_get_only_child_text_content("." + class_name)
+                for order_key, key in object_dict.items():
+                    index = int(order_key)
+                    if len(elements_texts) > index:
+                        return_json[key] = elements_texts[index]
+
+            class_multiObject_search_exist = config.get("class_multiObject_search_exist", {})
+            for class_name, object_list in class_multiObject_search_exist.items():
+                elements_texts = safely_get_text_content("." + class_name)
+                for each_object in object_list:
+                    if each_object == "is_other_exist":
+                        continue
+                    if each_object in elements_texts:
+                        return_json[each_object] = True
+                    else:
+                        return_json[each_object] = False
+                if "is_other_exist" in object_list:
+                    for each_element in elements_texts:
+                        if each_element not in object_list:
+                            return_json["is_other_exist"] = True
+                            break
+                    if "is_other_exist" not in return_json.keys():
+                        return_json["is_other_exist"] = False
+                

            class_singleObject = config.get("class_singleObject", {})
            for class_name, key in class_singleObject.items():
@@ -1240,6 +1301,55 @@ def get_active_tab_html_parse(env, config: Dict[str, Any]):
                inputs = target_page.locator(f"xpath={xpath}")
                if inputs.count() > 0:
                    return_json[key] = inputs.first.input_value().strip()
+        
+        elif config["category"] == "class&url":
+            class_multiObject = config.get("class_multiObject", {})
+            for class_name, object_list in class_multiObject.items():
+                elements_texts = safely_get_text_content("." + class_name)
+                for each_key in object_list:
+                    if any(each_key.lower() == text.lower() for text in elements_texts):
+                        return_json[each_key.lower()] = True
+
+                for each_key in elements_texts:
+                    # each_key.lower() not in object_list.lower():
+                    if all(each_key.lower() not in item.lower() for item in object_list):
+                        return_json["is_other_exist"] = True
+                        break
+                if "is_other_exist" not in return_json.keys():
+                    return_json["is_other_exist"] = False
+            
+            class_multiObject_li = config.get("class_multiObject_li", {})
+            for class_name, object_list in class_multiObject_li.items():
+                elements_texts = safely_get_direct_li_playwright("." + class_name)
+                for each_key in object_list:
+                    if any(each_key.lower() == text.lower() for text in elements_texts):
+                        return_json[each_key.lower()] = True
+                
+                for each_key in elements_texts:
+                    # each_key.lower() not in object_list.lower():
+                    if all(each_key.lower() not in item.lower() for item in object_list):
+                        return_json["is_other_exist"] = True
+                        break
+                if "is_other_exist" not in return_json.keys():
+                    return_json["is_other_exist"] = False
+
+            url_include_expected = config.get("url_include_expected", [])
+            for key in url_include_expected:
+                if key.lower() in target_page.url.lower():
+                    if key.lower() not in return_json.keys():
+                        return_json[key.lower()] = True
+                else:
+                    if key.lower() not in return_json.keys():
+                        return_json[key.lower()] = False
+            
+            url_include_expected_multichoice = config.get("url_include_expected_multichoice", {})
+            for key, value in url_include_expected_multichoice.items():
+                if key.lower() in target_page.url.lower():
+                    if value.lower() not in return_json.keys():
+                        return_json[value.lower()] = True
+                else:
+                    if value.lower() not in return_json.keys():
+                        return_json[value.lower()] = False

        browser.close()
    return return_json
@@ -1278,13 +1388,14 @@ def get_gotoRecreationPage_and_get_html_content(env, config: Dict[str, Any]):
            browser = p.chromium.connect_over_cdp(remote_debugging_url)
        page = browser.new_page()
        page.goto("https://www.recreation.gov/")
-        page.fill("input#hero-search-input", "Albion Basin")
+        page.fill("input#hero-search-input", "Diamond")
        page.click("button.nav-search-button")
        print("after first click")
-        time.sleep(2)
+        time.sleep(10)
        # Assuming .search-result-highlight--success leads to a new page or requires page load
        with page.expect_popup() as popup_info:
            page.click(".search-result-highlight--success")
+        time.sleep(30)
        print("after second click")
        newpage = popup_info.value
        newpage.wait_for_load_state()
@@ -1337,6 +1448,8 @@ def get_active_tab_url_parse(env, config: Dict[str, Any]):
            # change original key to new key, keep value unchange
            value = extracted_params.pop(key)
            extracted_params[config["replace"][key]] = value
+    if config.get("split_list", False):
+        extracted_params = {key: extracted_params[key].split(',') for key in extracted_params.keys()}
    return extracted_params


@@ -1366,3 +1479,57 @@ def get_url_dashPart(env, config: Dict[str, str]):
        return dash_part
    elif config["returnType"] == "json":
        return {config["key"]: dash_part}
+
+
+def get_url_path_parse(env, config: Dict[str, str]):
+    """
+    Parse Macy's product url path, extract:
+    - mens_clothing: true if 'mens-clothing' in path, else None
+    - t_shirts: true if any key 'Top_style' or 'Product_department' value is 'T-shirts', else None
+    - Men_regular_size_t, Price_discount_range (as list), Sleeve_length: as before, None if not found
+    All fields are None if not found for robustness.
+    """
+    from urllib.parse import urlparse, unquote
+    result = {}
+    # 1. Parse URL
+    active_tab_url = get_active_url_from_accessTree(env, config)
+    if active_tab_url is None:
+        return None
+    parsed = urlparse(active_tab_url)
+    path = unquote(parsed.path)
+    result = {}
+    # mens_clothing
+    result['mens_clothing'] = True if 'mens-clothing' in path else None
+    # key-value
+    path_parts = path.strip('/').split('/')
+    key_value_json = {}
+    tshirts_flag = False
+    if "mens-t-shirts" in path:
+        tshirts_flag = True
+    for i in range(len(path_parts)-1):
+        if ',' in path_parts[i] and ',' in path_parts[i+1]:
+            keys = [k.strip() for k in path_parts[i].split(',')]
+            values = [v.strip() for v in path_parts[i+1].split(',')]
+            for k, v in zip(keys, values):
+                if k == "Price_discount_range":
+                    key_value_json[k] = [item.strip() for item in v.split('|')] if v else None
+                else:
+                    key_value_json[k] = v if v else None
+                if (k == 'Top_style' or k == 'Product_department') and (v == 'T-shirts' or v == 'T-Shirts' or v == 'T-Shirt'):
+                    tshirts_flag = True
+            break
+    for field in ['Men_regular_size_t', 'Price_discount_range', 'Sleeve_length']:
+        if field not in key_value_json:
+            key_value_json[field] = None
+    result['t_shirts'] = tshirts_flag if tshirts_flag else None
+    # parse_keys
+    for key in config["parse_keys"]:
+        if key in key_value_json:
+            if key == "Price_discount_range":
+                if '50_PERCENT_ off & more' in key_value_json[key] and not '30_PERCENT_ off & more' in key_value_json[key] and not '20_PERCENT_ off & more' in key_value_json[key]:
+                    result[key] = '50_PERCENT_ off & more'
+                else:
+                    result[key] = 'not_50_PERCENT_ off & more'
+            else:
+                result[key] = key_value_json[key]
+    return result
--- a/desktop_env/evaluators/getters/misc.py
+++ b/desktop_env/evaluators/getters/misc.py
@@ -71,7 +71,10 @@ relativeTime_to_IntDay = {
    "this Sunday": "special",
    "next Monday": "special",
    "next Friday": "special",
-    "first monday four months later": "special"
+    "first monday four months later": "special",
+    "first monday eight months later": "special",
+    "next Monday split": "special",
+    "next Friday split": "special"
 }

 def get_rule(env, config: Dict[str, R]) -> R:
@@ -125,6 +128,12 @@ def get_rule_relativeTime(env, config: Dict[str, R]) -> R:
                # get the first monday of the next_month
                temp_date = datetime(next_year, next_month, 1)
                absoluteDay = temp_date + timedelta(days=((6-temp_date.weekday())+1)%7)
+            elif start_relative_time == "first monday eight months later":
+                next_year = now.year + 1 if now.month >= 5 else now.year
+                next_month = (now.month + 8)%12
+                # get the first monday of the next_month
+                temp_date = datetime(next_year, next_month, 1)
+                absoluteDay = temp_date + timedelta(days=((6-temp_date.weekday())+1)%7)
        regular_time = apply_rules_to_timeFormat(relativeRules["expected"]["time"], absoluteDay)
        config["rules"]["expected"]["time"] = regular_time

@@ -144,12 +153,20 @@ def get_rule_relativeTime(env, config: Dict[str, R]) -> R:
                next_month = now.month + 1 if now.month < 12 else 1
                next_day = 10
                from_absoluteDay = datetime(next_year, next_month, next_day)
-            elif from_time == "next Monday":
+            elif from_time == "next Monday" or from_time == "next Monday split":
               from_absoluteDay = now + timedelta(days=((6-now.weekday())+1)) 
            else:
                pass # more rules here
-        regular_from_time = apply_rules_to_timeFormat(relativeRules["expected"]["from"], from_absoluteDay)
-        config["rules"]["expected"]["from"] = regular_from_time
+        if from_time == "next Monday split":
+            puday = apply_rules_to_timeFormat(relativeRules["expected"]["puDay"], from_absoluteDay)
+            config["rules"]["expected"]["puDay"] = puday
+            pumonth = apply_rules_to_timeFormat(relativeRules["expected"]["puMonth"], from_absoluteDay)
+            config["rules"]["expected"]["puMonth"] = pumonth
+            puyear = apply_rules_to_timeFormat(relativeRules["expected"]["puYear"], from_absoluteDay)
+            config["rules"]["expected"]["puYear"] = puyear
+        else:
+            regular_from_time = apply_rules_to_timeFormat(relativeRules["expected"]["from"], from_absoluteDay)
+            config["rules"]["expected"]["from"] = regular_from_time

        # deal with to_time
        if relativeTime_to_IntDay[to_time] != "special":
@@ -164,15 +181,23 @@ def get_rule_relativeTime(env, config: Dict[str, R]) -> R:
                next_month = now.month + 1 if now.month < 12 else 1
                next_day = 11
                to_absoluteDay = datetime(next_year, next_month, next_day)
-            elif to_time == "next Friday":
+            elif to_time == "next Friday" or to_time == "next Friday split":
                if now.weekday() < 4 and from_time in ["next Monday"]:
                    to_absoluteDay = now + timedelta(days=((4-now.weekday())+7))
                else:
                    to_absoluteDay = now + timedelta(days=((4-now.weekday()) if now.weekday() < 4 else (6-now.weekday()) + 5))
            else:
                pass # more rules here
-        regular_to_time = apply_rules_to_timeFormat(relativeRules["expected"]["to"], to_absoluteDay)
-        config["rules"]["expected"]["to"] = regular_to_time
+        if to_time == "next Friday split":
+            to_day = apply_rules_to_timeFormat(relativeRules["expected"]["doDay"], to_absoluteDay)
+            config["rules"]["expected"]["doDay"] = to_day
+            to_month = apply_rules_to_timeFormat(relativeRules["expected"]["doMonth"], to_absoluteDay)
+            config["rules"]["expected"]["doMonth"] = to_month
+            to_year = apply_rules_to_timeFormat(relativeRules["expected"]["doYear"], to_absoluteDay)
+            config["rules"]["expected"]["doYear"] = to_year
+        else:
+            regular_to_time = apply_rules_to_timeFormat(relativeRules["expected"]["to"], to_absoluteDay)
+            config["rules"]["expected"]["to"] = regular_to_time
            
    return config["rules"]

@@ -186,6 +211,7 @@ def apply_rules_to_timeFormat(timeFormat: str, absoluteDay: datetime):
    timeFormat = timeFormat.replace("{month}", month_mapping_full[absoluteDay.month], 1)
    timeFormat = timeFormat.replace("{MonthFull}", Month_Mapping_Full[absoluteDay.month], 1)
    timeFormat = timeFormat.replace("{Day0D}", "0"+str(absoluteDay.day) if absoluteDay.day < 10 else str(absoluteDay.day), 1)
+    timeFormat = timeFormat.replace("{MonthD}", str(absoluteDay.month), 1)
    # you can add other replace rules here

    return timeFormat
--- a/desktop_env/evaluators/metrics/init.py
+++ b/desktop_env/evaluators/metrics/init.py
@@ -21,7 +21,8 @@ from .chrome import (
    is_expected_url_pattern_match,
    is_added_to_steam_cart,
    is_expected_installed_extensions,
-    compare_pdf_images
+    compare_pdf_images,
+    is_expected_active_tab_approximate
 )
 from .docs import (
    compare_font_names,
--- a/desktop_env/evaluators/metrics/chrome.py
+++ b/desktop_env/evaluators/metrics/chrome.py
@@ -37,6 +37,34 @@ def is_expected_active_tab(active_tab_info: Dict[str, str], rule: Dict[str, Any]
        return 0


+def is_expected_active_tab_approximate(active_tab_info: Dict[str, str], rule: Dict[str, Any]) -> float:
+    """
+    Checks if the expected active tab is open in Chrome, ignoring query parameters in the URL.
+    """
+    if not active_tab_info:
+        return 0.
+
+    match_type = rule['type']
+
+    if match_type == "url":
+        expected_url = rule['url']
+        if isinstance(active_tab_info, Dict):
+            actual_url = active_tab_info.get('url', None)
+        else:
+            actual_url = active_tab_info
+        from urllib.parse import urlparse, urlunparse
+        def strip_query(url):
+            parsed = urlparse(url)
+            return urlunparse(parsed._replace(query=""))
+        if strip_query(expected_url) == strip_query(actual_url):
+            return 1
+        else:
+            return 0
+    else:
+        logger.error(f"Unknown type: {match_type}")
+        return 0
+
+
 # rules[expected] is a string-formatted regex
 def is_expected_url_pattern_match(result, rules) -> float:
    """
@@ -366,7 +394,12 @@ def is_shortcut_on_desktop(shortcuts: Dict[str, str], rule):
        for shortcut_path, shortcut_content in shortcuts.items():
            if "Name=" + rule['name'] + "\n" in shortcut_content:
                return 1.
-        return 0.
+        return 0.0
+    elif rule['type'] == 'exec':
+        for shortcut_path, shortcut_content in shortcuts.items():
+            if "Exec=" + rule['exec'] + "\n" in shortcut_content:
+                return 1.
+        return 0.0
    elif rule['type'] == 'url':
        raise TypeError(f"{rule['type']} not support yet!")
    elif rule['type'] == 'id':
--- a/desktop_env/evaluators/metrics/general.py
+++ b/desktop_env/evaluators/metrics/general.py
@@ -322,8 +322,14 @@ def check_direct_json_object(result, rules) -> float:
            expected_json = rules["expected"]
            for key in expected_json.keys():
                expected_value = expected_json.get(key)
-                if expected_value != result.get(key):
-                    return 0.
+                if expected_json.get("ignore_list_order", False):
+                    expected_value = sorted(expected_value)
+                    result_value = sorted(result.get(key))
+                    if expected_value != result_value:
+                        return 0.
+                else:
+                    if expected_value != result.get(key):
+                        return 0.
            return 1.0
        else:
            expected_json = rules["expected"]
--- a/evaluation_examples/examples/chrome/12086550-11c0-466b-b367-1d9e75b3910e.json
+++ b/evaluation_examples/examples/chrome/12086550-11c0-466b-b367-1d9e75b3910e.json
@@ -29,7 +29,7 @@
    "chrome"
  ],
  "evaluator": {
-    "func": "is_expected_active_tab",
+    "func": "is_expected_active_tab_approximate",
    "result": {
      "type": "active_url_from_accessTree",
      "goto_prefix": ""
--- a/evaluation_examples/examples/chrome/1704f00f-79e6-43a7-961b-cedd3724d5fd.json
+++ b/evaluation_examples/examples/chrome/1704f00f-79e6-43a7-961b-cedd3724d5fd.json
@@ -1,7 +1,7 @@
 {
  "id": "1704f00f-79e6-43a7-961b-cedd3724d5fd",
  "snapshot": "chrome",
-  "instruction": "Find a large car with lowest price from next Monday to next Friday in Zurich.",
+  "instruction": "Find a large car from next Monday to Friday in Zurich, sorted by price.",
  "source": "test_task_0",
  "config": [
    {
@@ -59,13 +59,16 @@
        ]
      },
      {
-        "type": "active_tab_html_parse",
+        "type": "active_tab_url_parse",
        "goto_prefix": "https://www.",
-        "category": "xpath",
-        "xpathObject": {
-          "/html/body/main/div/div/div/section/div/div/div/div[1]/div[1]/p": "from",
-          "/html/body/main/div/div/div/section/div/div/div/div[1]/div[3]/p": "to"
-        }
+        "parse_keys": [
+          "puDay",
+          "puMonth",
+          "puYear",
+          "doDay",
+          "doMonth",
+          "doYear"
+        ]
      }
    ],
    "expected": [
@@ -77,6 +80,7 @@
            "dropLocationName": "Zürich",
            "filterCriteria_carCategory": "large",
            "filterCriteria_sortBy": "PRICE"
+
          }
        }
      },
@@ -84,12 +88,16 @@
        "type": "rule_relativeTime",
        "rules": {
          "relativeTime": {
-            "from": "next Monday",
-            "to": "next Friday"
+            "from": "next Monday split",
+            "to": "next Friday split"
          },
          "expected": {
-            "from": "{DoW}, {DayD} {Month} {Year}, 10:00",
-            "to": "{DoW}, {DayD} {Month} {Year}, 10:00"
+            "puDay": "{DayD}",
+            "puMonth": "{MonthD}",
+            "puYear": "{Year}",
+            "doDay": "{DayD}",
+            "doMonth": "{MonthD}",
+            "doYear":"{Year}"
          }
        }
      }
--- a/evaluation_examples/examples/chrome/2888b4e6-5b47-4b57-8bf5-c73827890774.json
+++ b/evaluation_examples/examples/chrome/2888b4e6-5b47-4b57-8bf5-c73827890774.json
@@ -1,7 +1,7 @@
 {
  "id": "2888b4e6-5b47-4b57-8bf5-c73827890774",
  "snapshot": "chrome",
-  "instruction": "Find a men's T-Shirt that is in large size with a stripe pattern, short sleeve and under the Sales&Discount.",
+  "instruction": "Show me all men's large-size short-sleeve T-shirts with a discount of 50% or more.",
  "source": "test_task_1",
  "config": [
    {
@@ -43,18 +43,28 @@
    "chrome"
  ],
  "evaluator": {
-    "func": "exact_match",
+    "func": "check_direct_json_object",
    "result": {
-      "type": "url_dashPart",
+      "type": "url_path_parse",
      "goto_prefix": "https://www.",
-      "partIndex": -1,
-      "needDeleteId": true,
-      "returnType": "string"
+      "parse_keys": [
+        "mens_clothing",
+        "t_shirts",
+        "Men_regular_size_t",
+        "Price_discount_range",
+        "Sleeve_length"
+      ]
    },
    "expected": {
      "type": "rule",
      "rules": {
-        "expected": "Stripe,Men,L,Short%20Sleeve,Sales%20%26%20Discounts"
+        "expected": {
+          "mens_clothing": true,
+          "t_shirts": true,
+          "Men_regular_size_t": "L",
+          "Price_discount_range": "50_PERCENT_ off & more",
+          "Sleeve_length": "Short Sleeve"
+        }
      }
    }
  },
--- a/evaluation_examples/examples/chrome/35253b65-1c19-4304-8aa4-6884b8218fc0.json
+++ b/evaluation_examples/examples/chrome/35253b65-1c19-4304-8aa4-6884b8218fc0.json
@@ -44,8 +44,8 @@
    "expected": {
      "type": "rule",
      "rules": {
-        "type": "name",
-        "name": "Play Puzzle Game 2048"
+        "type": "exec",
+        "exec": "/opt/google/chrome/google-chrome --profile-directory=Default --app-id=poahllcmmahlafplfhgjomkjmeblpapf"
      }
    }
  },
--- a/evaluation_examples/examples/chrome/47543840-672a-467d-80df-8f7c3b9788c9.json
+++ b/evaluation_examples/examples/chrome/47543840-672a-467d-80df-8f7c3b9788c9.json
@@ -1,7 +1,7 @@
 {
  "id": "47543840-672a-467d-80df-8f7c3b9788c9",
  "snapshot": "chrome",
-  "instruction": "Find and select the car with the most number of seats to pick up in Boston Logan Intl Airport from 10th next month to 11th next month.",
+  "instruction": "Show me the cars available for pickup at Boston Logan Intl Airport from the 10th to the 11th of next month, sorted by the number of seats to find the largest capacity.",
  "source": "test_task_1",
  "config": [
    {
@@ -75,7 +75,7 @@
        "goto_prefix": "https://www.",
        "category": "xpath",
        "xpathObject": {
-          "/html/body/div[6]/div[2]/div[1]/div/div/div[2]/div[1]/section[1]/div/form/div[1]/div[2]/div/a": "rank"
+          "/html/body/div[6]/div[2]/div[1]/div/div/div[2]/section[1]/div[1]/form/div[1]/div[1]/div[2]/div/a": "rank"
        }
      }
    ],
--- a/evaluation_examples/examples/chrome/6766f2b8-8a72-417f-a9e5-56fcaa735837.json
+++ b/evaluation_examples/examples/chrome/6766f2b8-8a72-417f-a9e5-56fcaa735837.json
@@ -1,7 +1,7 @@
 {
  "id": "6766f2b8-8a72-417f-a9e5-56fcaa735837",
  "snapshot": "chrome",
-  "instruction": "Could you help me install the unpacked extension at /home/user/Desktop/ in Chrome?",
+  "instruction": "Could you help me unzip the downloaded extension file from /home/user/Desktop/ to /home/user/Desktop/ and configure it in Chrome's extensions?",
  "source": "https://support.google.com/chrome/thread/205881926/it-s-possible-to-load-unpacked-extension-automatically-in-chrome?hl=en",
  "config": [
    {
--- a/evaluation_examples/examples/chrome/6c4c23a1-42a4-43cc-9db1-2f86ff3738cc.json
+++ b/evaluation_examples/examples/chrome/6c4c23a1-42a4-43cc-9db1-2f86ff3738cc.json
@@ -49,11 +49,11 @@
      "goto_prefix": "https://www.",
      "category": "class",
      "class_singleObject": {
-        "search-date": "time",
-        "price-in-tabs__nav--selected": "category"
+        "mach-flight-context-info__wrapper--date": "time",
+        "mach-global-tabs-small__wrapper__tab--active": "category"
      },
-      "class_multiObject": {
-        "search-segment-cities__city": {
+      "class_multiObject_child": {
+        "mach-flight-context-info__wrapper__info--separator": {
          "0": "start",
          "1": "end"
        }
@@ -68,7 +68,7 @@
        "expected": {
          "start": "SEA",
          "end": "NYC",
-          "time": "{DoW}, {Month} {DayD}, {Year}",
+          "time": "{DoW}, {Month} {Day0D}, {Year}",
          "category": "Miles"
        }
      }
--- a/evaluation_examples/examples/chrome/7f52cab9-535c-4835-ac8c-391ee64dc930.json
+++ b/evaluation_examples/examples/chrome/7f52cab9-535c-4835-ac8c-391ee64dc930.json
@@ -43,24 +43,52 @@
    "chrome"
  ],
  "evaluator": {
-    "func": "check_direct_json_object",
-    "result": {
-      "type": "active_tab_url_parse",
-      "goto_prefix": "https://www.",
-      "parse_keys": [
-        "q",
-        "tbs"
-      ]
-    },
-    "expected": {
-      "type": "rule",
-      "rules": {
-        "expected": {
-          "q": "drip coffee maker",
-          "tbs": "mr:1,price:1,ppr_min:25,ppr_max:60,sales:1,pdtr0:1825161|1825162"
+    "func": [
+      "check_direct_json_object",
+      "check_direct_json_object"
+    ],
+    "result": [
+      {
+        "type": "active_tab_url_parse",
+        "goto_prefix": "https://www.",
+        "parse_keys": [
+          "q"
+        ]
+      },
+      {
+        "type": "active_tab_html_parse",
+        "goto_prefix": "https://www.",
+        "category": "class",
+        "class_multiObject_search_exist": {
+          "fT28tf":[
+            "Black",
+            "$25 - $60",
+            "is_other_exist"
+          ]
        }
      }
-    }
+    ],
+    "expected": [
+      {
+        "type": "rule",
+        "rules": {
+          "expected": {
+            "q": "drip coffee maker"
+          },
+          "expect_in_result": true
+        }
+      },
+      {
+        "type": "rule",
+        "rules": {
+          "expected": {
+            "Black": true,
+            "$25 - $60": true,
+            "is_other_exist": false
+          }
+        }
+      }
+    ]
  },
  "proxy": true
 }
--- a/evaluation_examples/examples/chrome/9f3f70fc-5afc-4958-a7b7-3bb4fcb01805.json
+++ b/evaluation_examples/examples/chrome/9f3f70fc-5afc-4958-a7b7-3bb4fcb01805.json
@@ -43,41 +43,37 @@
    "chrome"
  ],
  "evaluator": {
-    "func": [
-      "is_expected_url_pattern_match",
-      "check_direct_json_object"
-    ],
-    "conj": "and",
-    "result": [
-      {
-        "type": "active_tab_info"
+    "func": "check_direct_json_object",
+    "result": {
+      "type": "active_tab_html_parse",
+      "category": "class&url",
+      "class_multiObject": {
+        "filter-selector-link": [
+          "over $60",
+          "women",
+          "jerseys",
+          "nike"
+        ]
      },
-      {
-        "type": "active_tab_html_parse",
-        "category": "xpath",
-        "xpathObject": {
-          "/html/body/div[2]/div/div[6]/div[2]/div[2]/div/div[1]/div[4]/ul/li[2]": "money"
+      "url_include_expected": [
+        "over $60",
+        "women",
+        "jerseys",
+        "nike"
+      ]
+    },
+    "expected": {
+      "type": "rule",
+      "rules": {
+        "expected": {
+          "over $60": true,
+          "women": true,
+          "jerseys": true,
+          "nike": true,
+          "is_other_exist": false
        }
      }
-    ],
-    "expected": [
-      {
-        "type": "rule",
-        "rules": {
-          "expected": [
-            "/women-jerseys/"
-          ]
-        }
-      },
-      {
-        "type": "rule",
-        "rules": {
-          "expected": {
-            "money": "over $60"
-          }
-        }
-      }
-    ]
+    }
  },
  "proxy": true
 }
--- a/evaluation_examples/examples/chrome/b4f95342-463e-4179-8c3f-193cd7241fb2.json
+++ b/evaluation_examples/examples/chrome/b4f95342-463e-4179-8c3f-193cd7241fb2.json
@@ -1,7 +1,7 @@
 {
  "id": "b4f95342-463e-4179-8c3f-193cd7241fb2",
  "snapshot": "chrome",
-  "instruction": "Find the next available date for Albion Basin.",
+  "instruction": "Find the next available date for Diamond.",
  "source": "test_task_1",
  "config": [
    {
--- a/evaluation_examples/examples/chrome/b7895e80-f4d1-4648-bee0-4eb45a6f1fa8.json
+++ b/evaluation_examples/examples/chrome/b7895e80-f4d1-4648-bee0-4eb45a6f1fa8.json
@@ -49,11 +49,11 @@
      "goto_prefix": "https://www.",
      "category": "xpath",
      "xpathObject": {
-        "/html/body/div[1]/main/div[3]/div/div[1]/div[2]/div[1]/div[2]/div/div/div/div/div[1]/div/button/div[3]": "from",
-        "/html/body/div[1]/main/div[3]/div/div[1]/div[2]/div[1]/div[2]/div/div/div/div/div[2]/button/div[3]": "to",
-        "/html/body/div[1]/main/div[3]/div/div[1]/div[2]/div[1]/div[1]/div/h1": "city",
-        "/html/body/div[1]/main/div[3]/div/div[1]/div[2]/div[1]/div[2]/div/div/div/div/div[3]/button/div[3]/span/span[2]": "adult",
-        "/html/body/div[1]/main/div[3]/div/div[2]/div/div[1]/div/div[2]/div[1]/div/div[1]/div/div[1]/div[2]/div/div[2]/div/button/div/div": "rank"
+        "/html/body/div[1]/main/div[3]/div[5]/div[2]/div/div[1]/div/div/div/div[1]/div/button/div[3]": "from",
+        "/html/body/div[1]/main/div[3]/div[5]/div[2]/div/div[1]/div/div/div/div[2]/button/div[3]": "to",
+        "/html/body/div[1]/main/div[3]/div[2]/div/div[1]/div/h2": "city",
+        "/html/body/div[1]/main/div[3]/div[5]/div[2]/div/div[1]/div/div/div/div[3]/button/div[3]/span/span[2]": "adult",
+        "/html/body/div[1]/main/div[3]/div[5]/div[2]/div/div[3]/div/div[2]/div/div/div[2]/div/button/div/div": "rank"
      }
    },
    "expected": {
@@ -67,7 +67,7 @@
          "from": "{DoW}, {Month} {Day0D}",
          "to": "{DoW}, {Month} {Day0D}",
          "city": "New York City Hotels",
-          "adult": "2 adults",
+          "adult": "2 guests",
          "rank": "Price (low to high)"
        }
      }
--- a/evaluation_examples/examples/chrome/c1fa57f3-c3db-4596-8f09-020701085416.json
+++ b/evaluation_examples/examples/chrome/c1fa57f3-c3db-4596-8f09-020701085416.json
@@ -1,7 +1,7 @@
 {
  "id": "c1fa57f3-c3db-4596-8f09-020701085416",
  "snapshot": "chrome",
-  "instruction": "Open the baggage fee calculator.",
+  "instruction": "Open the baggage fee calculator in United Airlines website.",
  "source": "test_task_1",
  "config": [
    {
@@ -62,7 +62,7 @@
      "type": "rule",
      "rules": {
        "expected": [
-          "checked-bag-fee-calculator"
+          "united.com/en/us/checked-bag-fee-calculator"
        ]
      }
    }
--- a/evaluation_examples/examples/chrome/cabb3bae-cccb-41bd-9f5d-0f3a9fecd825.json
+++ b/evaluation_examples/examples/chrome/cabb3bae-cccb-41bd-9f5d-0f3a9fecd825.json
@@ -43,19 +43,39 @@
    "chrome"
  ],
  "evaluator": {
-    "func": "is_expected_url_pattern_match",
+    "func": "check_direct_json_object",
    "result": {
-      "type": "active_url_from_accessTree",
-      "goto_prefix": "https://www."
+      "type": "active_tab_html_parse",
+      "goto_prefix": "https://www.",
+      "category": "class&url",
+      "class_multiObject_li": {
+        "pmpSearch_breadcrumb": [
+          "Spider-Man",
+          "Toys",
+          "Kids"
+        ],
+        "sbSelector": [
+          "Price Low-High"
+        ]
+      },
+      "url_include_expected_multichoice": {
+        "Spider-Man": "Spider-Man",
+        "spiderman": "Spider-Man",
+        "Toys": "Toys",
+        "Kids": "Kids",
+        "S=4": "Price Low-High"
+      }
    },
    "expected": {
      "type": "rule",
      "rules": {
-        "expected": [
-          "AgeAppropriate:Kids",
-          "search=spider[-%20]?man%20toys",
-          "S=4"
-        ]
+        "expected": {
+          "spider-man": true,
+          "toys": true,
+          "kids": true,
+          "price low-high": true,
+          "is_other_exist": false
+        }
      }
    }
  },
--- a/evaluation_examples/examples/chrome/da46d875-6b82-4681-9284-653b0c7ae241.json
+++ b/evaluation_examples/examples/chrome/da46d875-6b82-4681-9284-653b0c7ae241.json
@@ -1,7 +1,7 @@
 {
  "id": "da46d875-6b82-4681-9284-653b0c7ae241",
  "snapshot": "chrome",
-  "instruction": "Schedule an appointment to apply for transportation access pass in  the Charlie Card store on the first Monday four months later, 10:15 am, fill in my details (James Smith, james.smith@gmail.com). And don not click \"book\" directly. Let me review it.",
+  "instruction": "Book an appointment to apply for a transportation access pass at the Charlie Card store on the first Monday eight months later, 10:15 am, fill in my details (James Smith, james.smith@gmail.com). And don not click \"book\" directly. Let me review it.",
  "source": "test_task_2",
  "config": [
    {
@@ -56,11 +56,10 @@
      {
        "type": "active_tab_html_parse",
        "category": "class",
-        "class_singleObject": {},
-        "class_multiObject": {
-          "breakword": {
-            "1": "content",
-            "2": "time"
+        "class_multiObject_only_child": {
+          "HAZ16": {
+            "0": "content",
+            "1": "time"
          }
        }
      },
@@ -68,8 +67,8 @@
        "type": "active_tab_html_parse",
        "category": "input",
        "inputObject": {
-          "/html/body/div/div/form/div[7]/div/div/div[1]/input[1]": "name",
-          "/html/body/div/div/form/div[7]/div/div/div[1]/input[2]": "mail"
+          "/html/body/div[2]/div/form/div[7]/div/div/div[1]/input[1]": "name",
+          "/html/body/div[2]/div/form/div[7]/div/div/div[1]/input[2]": "mail"
        }
      }
    ],
@@ -78,7 +77,7 @@
        "type": "rule",
        "rules": {
          "expected": [
-            "CharlieCardStoreAppointments@mbta.com/bookings/"
+            "book/CharlieCardStoreAppointments@mbta.com/"
          ]
        }
      },
@@ -86,11 +85,11 @@
        "type": "rule_relativeTime",
        "rules": {
          "relativeTime": {
-            "from": "first monday four months later"
+            "from": "first monday eight months later"
          },
          "expected": {
            "content": "Apply for Transportation Access Pass (TAP) CharlieCard non-auto approval",
-            "time": "{MonthFull} {Day0D}, 10:15 am"
+            "time": "{MonthFull} {Day0D}, 10:15 AM"
          }
        }
      },
--- a/evaluation_examples/examples/chrome/e1e75309-3ddb-4d09-92ec-de869c928143.json
+++ b/evaluation_examples/examples/chrome/e1e75309-3ddb-4d09-92ec-de869c928143.json
@@ -1,7 +1,7 @@
 {
  "id": "e1e75309-3ddb-4d09-92ec-de869c928143",
  "snapshot": "chrome",
-  "instruction": "Computer, can you turn the webpage I'm looking at into a PDF file and put it on my main screen, you know, the Desktop?",
+  "instruction": "Computer, can you turn the webpage I'm looking at into a PDF file, save it to my Desktop with the default filename and set the margins to none?",
  "source": "https://in5stepstutorials.com/google-chrome/save-web-page-as-pdf-in-chrome.php",
  "config": [
    {
--- a/evaluation_examples/examples/chrome/f0b971a1-6831-4b9b-a50e-22a6e47f45ba.json
+++ b/evaluation_examples/examples/chrome/f0b971a1-6831-4b9b-a50e-22a6e47f45ba.json
@@ -1,7 +1,7 @@
 {
  "id": "f0b971a1-6831-4b9b-a50e-22a6e47f45ba",
  "snapshot": "chrome",
-  "instruction": "Show me the scores for the 2019 super bowl",
+  "instruction": "Please help me find the score record for the 2019 Super Bowl in the NFL website.",
  "source": "Mind2Web",
  "config": [
    {
@@ -62,7 +62,7 @@
      "type": "rule",
      "rules": {
        "type": "url",
-        "url": "https://www.nfl.com/scores/2019/POST4"
+        "url": "https://www.nfl.com/scores/2019/post4"
      }
    }
  },
--- a/evaluation_examples/examples/chrome/f3b19d1e-2d48-44e9-b4e1-defcae1a0197.json
+++ b/evaluation_examples/examples/chrome/f3b19d1e-2d48-44e9-b4e1-defcae1a0197.json
@@ -1,7 +1,7 @@
 {
  "id": "f3b19d1e-2d48-44e9-b4e1-defcae1a0197",
  "snapshot": "chrome",
-  "instruction": "Find help page about buying tickets.",
+  "instruction": "Find the FAQ page about ticket delivery.",
  "source": "test_task_0",
  "config": [
    {
@@ -27,7 +27,7 @@
      "type": "chrome_open_tabs",
      "parameters": {
        "urls_to_open": [
-          "https://seatgeek.com/"
+          "https://premier.ticketek.com.au/"
        ]
      }
    },
@@ -45,17 +45,16 @@
  "evaluator": {
    "func": "is_expected_url_pattern_match",
    "result": {
-      "type": "active_tab_info",
-      "goto_prefix": "https://www."
+      "type": "active_tab_info"
    },
    "expected": {
      "type": "rule",
      "rules": {
        "expected": [
-          "Buying-Tickets"
+          "Ticket-Delivery-FAQs"
        ]
      }
    }
  },
-  "proxy": true
+  "proxy": false
 }
--- a/evaluation_examples/examples/chrome/f5d96daf-83a8-4c86-9686-bada31fc66ab.json
+++ b/evaluation_examples/examples/chrome/f5d96daf-83a8-4c86-9686-bada31fc66ab.json
@@ -1,7 +1,7 @@
 {
  "id": "f5d96daf-83a8-4c86-9686-bada31fc66ab",
  "snapshot": "chrome",
-  "instruction": "Compare iPhone 15 Pro Max with iPhone 13 Pro Max",
+  "instruction": "Compare iPhone 15 Pro Max with iPhone 14 Pro Max and iPhone 13 Pro Max",
  "source": "Mind2Web",
  "config": [
    {
@@ -53,16 +53,26 @@
    "chrome"
  ],
  "evaluator": {
-    "func": "is_expected_active_tab",
+    "func": "check_direct_json_object",
    "result": {
-      "type": "active_url_from_accessTree",
-      "goto_prefix": "https://www."
+      "type": "active_tab_url_parse",
+      "goto_prefix": "https://www.",
+      "parse_keys": [
+        "modelList"
+      ],
+      "split_list": true
    },
    "expected": {
      "type": "rule",
      "rules": {
-        "type": "url",
-        "url": "https://www.apple.com/iphone/compare/?modelList=iphone-15-pro-max,iphone-15-pro,iphone-13-pro-max"
+        "expected": {
+          "modelList": [
+            "iphone-15-pro-max",
+            "iphone-14-pro-max",
+            "iphone-13-pro-max"
+          ]
+        },
+        "ignore_list_order": true
      }
    }
  },
--- a/evaluation_examples/examples/chrome/fc6d8143-9452-4171-9459-7f515143419a.json
+++ b/evaluation_examples/examples/chrome/fc6d8143-9452-4171-9459-7f515143419a.json
@@ -49,10 +49,10 @@
      "goto_prefix": "https://www.",
      "category": "class",
      "class_singleObject": {
-        "search-date": "time"
+        "mach-flight-context-info__wrapper--date": "time"
      },
-      "class_multiObject": {
-        "search-segment-cities__city": {
+      "class_multiObject_child": {
+        "mach-flight-context-info__wrapper__info--separator": {
          "0": "start",
          "1": "end"
        }
@@ -67,7 +67,7 @@
        "expected": {
          "start": "NYC",
          "end": "CMH",
-          "time": "{DoW}, {Month} {DayD}, {Year}"
+          "time": "{DoW}, {Month} {Day0D}, {Year}"
        }
      }
    }