datamarket 0.9.37__tar.gz → 0.9.39__tar.gz

{datamarket-0.9.37 → datamarket-0.9.39}/PKG-INFO

@@ -1,8 +1,7 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.3
 Name: datamarket
-Version: 0.9.37
+Version: 0.9.39
 Summary: Utilities that integrate advanced scraping knowledge into just one library.
-Home-page: https://datamarket.es
 License: GPL-3.0-or-later
 Author: DataMarket
 Author-email: techsupport@datamarket.es
@@ -37,6 +36,7 @@ Provides-Extra: json5
 Provides-Extra: lxml
 Provides-Extra: nodriver
 Provides-Extra: openpyxl
+Provides-Extra: pandarallel
 Provides-Extra: pandas
 Provides-Extra: pandera
 Provides-Extra: peerdb
@@ -91,11 +91,12 @@ Requires-Dist: lxml[html-clean] (>=5.0.0,<6.0.0) ; extra == "lxml"
 Requires-Dist: nodriver (>=0.44,<0.45) ; extra == "nodriver"
 Requires-Dist: numpy (>=2.0.0,<3.0.0)
 Requires-Dist: openpyxl (>=3.0.0,<4.0.0) ; extra == "openpyxl"
+Requires-Dist: pandarallel (>=1.0.0,<2.0.0) ; extra == "pandarallel"
 Requires-Dist: pandas (>=2.0.0,<3.0.0) ; extra == "pandas"
 Requires-Dist: pandera (>=0.22.0,<0.23.0) ; extra == "pandera"
 Requires-Dist: pendulum (>=3.0.0,<4.0.0)
 Requires-Dist: pillow (>=11.0.0,<12.0.0) ; extra == "pillow"
-Requires-Dist: playwright (==1.47.0) ; extra == "playwright"
+Requires-Dist: playwright (==1.47.0) ; extra == "playwright" or extra == "camoufox"
 Requires-Dist: pre-commit (>=4.0.0,<5.0.0)
 Requires-Dist: presidio-analyzer[phonenumbers] (>=2.0.0,<3.0.0) ; extra == "pii"
 Requires-Dist: presidio-anonymizer (>=2.0.0,<3.0.0) ; extra == "pii"
@@ -112,18 +113,18 @@ Requires-Dist: rapidfuzz (>=3.0.0,<4.0.0) ; extra == "rapidfuzz"
 Requires-Dist: requests (>=2.0.0,<3.0.0)
 Requires-Dist: retry (>=0.9.0,<0.10.0) ; extra == "retry"
 Requires-Dist: shapely (>=2.0.0,<3.0.0) ; extra == "shapely"
-Requires-Dist: soda-core-mysql (>=3.0.0,<4.0.0) ; extra == "soda-core-mysql"
+Requires-Dist: soda-core-mysql-utf8-hotfix (>=3.0.0,<4.0.0) ; extra == "soda-core-mysql"
 Requires-Dist: soda-core-postgres (>=3.0.0,<4.0.0) ; extra == "soda-core-postgres"
 Requires-Dist: spacy (>=3.0.0,<4.0.0) ; extra == "pii"
 Requires-Dist: spacy-langdetect (>=0.1.0,<0.2.0) ; extra == "pii"
 Requires-Dist: stem (>=1.0.0,<2.0.0) ; extra == "stem" or extra == "proxy"
 Requires-Dist: tenacity (>=9.0.0,<10.0.0)
-Requires-Dist: tf-playwright-stealth (>=1.0.0,<2.0.0)
 Requires-Dist: tqdm (>=4.0.0,<5.0.0) ; extra == "tqdm"
 Requires-Dist: typer (>=0.15.0,<0.16.0)
 Requires-Dist: unidecode (>=1.0.0,<2.0.0)
 Requires-Dist: xmltodict (>=0.14.0,<0.15.0) ; extra == "xmltodict"
 Project-URL: Documentation, https://github.com/Data-Market/datamarket
+Project-URL: Homepage, https://datamarket.es
 Project-URL: Repository, https://github.com/Data-Market/datamarket
 Description-Content-Type: text/markdown
 

{datamarket-0.9.37 → datamarket-0.9.39}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "datamarket"
-version = "0.9.37"
+version = "0.9.39"
 description = "Utilities that integrate advanced scraping knowledge into just one library."
 authors = ["DataMarket <techsupport@datamarket.es>"]
 license = "GPL-3.0-or-later"
@@ -43,7 +43,7 @@ pytest = { version = "^8.0.0", optional = true }
 playwright = { version = "1.47.0", optional = true }
 tf-playwright-stealth = { version = "^1.0.0", optional = true }
 soda-core-postgres = { version = "^3.0.0", optional = true }
-soda-core-mysql = { version = "^3.0.0", optional = true }
+soda-core-mysql-utf8-hotfix = { version = "^3.0.0", optional = true }
 fake-useragent = { version = "^2.0.0", optional = true }
 pydrive2 = { version = "^1.0.0", optional = true }
 clickhouse-driver = { version = "~0.2.0", optional = true }
@@ -82,6 +82,7 @@ presidio-analyzer = { version = "^2.0.0", optional = true, extras = [
 presidio-anonymizer = { version = "^2.0.0", optional = true }
 spacy = { version = "^3.0.0", optional = true }
 spacy-langdetect = { version = "~0.1.0", optional = true }
+pandarallel = { version = "^1.0.0", optional = true }
 
 [tool.poetry.extras]
 boto3 = ["boto3"]
@@ -93,7 +94,7 @@ pytest = ["pytest"]
 playwright = ["playwright"]
 playwright-stealth = ["playwright-stealth"]
 soda-core-postgres = ["soda-core-postgres"]
-soda-core-mysql = ["soda-core-mysql"]
+soda-core-mysql = ["soda-core-mysql-utf8-hotfix"]
 fake-useragent = ["fake-useragent"]
 pydrive2 = ["pydrive2"]
 clickhouse-driver = ["clickhouse-driver"]
@@ -125,7 +126,8 @@ google-auth-oauthlib = ["google-auth-oauthlib"]
 dnspython = ["dnspython"]
 openpyxl = ["openpyxl"]
 httpx = ["httpx"]
-camoufox = ["camoufox"]
+camoufox = ["camoufox", "browserforge", "playwright"]
+pandarallel = ["pandarallel"]
 
 # Interface groups
 aws = ["boto3"]

datamarket-0.9.39/src/datamarket/utils/playwright/async_api.py

@@ -0,0 +1,23 @@
+########################################################################################################################
+# IMPORTS
+
+import asyncio
+from random import randint
+
+from playwright.async_api import Page
+
+
+########################################################################################################################
+# FUNCTIONS
+
+
+async def human_type(page: Page, text: str, delay: int = 100):
+    for char in text:
+        await page.keyboard.type(char, delay=randint(int(delay * 0.5), int(delay * 1.5)))  # noqa: S311
+
+
+async def human_press_key(page: Page, key: str, count: int = 1, delay: int = 100, sleep=True):
+    for _ in range(count):
+        await page.keyboard.press(key, delay=randint(int(delay * 0.5), int(delay * 1.5)))  # noqa: S311
+        if sleep:
+            await asyncio.sleep(randint(int(delay * 1.5), int(delay * 2.5)) / 1000)  # noqa: S311

datamarket-0.9.39/src/datamarket/utils/playwright/sync_api.py

@@ -0,0 +1,23 @@
+########################################################################################################################
+# IMPORTS
+
+import time
+from random import randint
+
+from playwright.sync_api import Page
+
+
+########################################################################################################################
+# FUNCTIONS
+
+
+def human_type(page: Page, text: str, delay: int = 100):
+    for char in text:
+        page.keyboard.type(char, delay=randint(int(delay * 0.5), int(delay * 1.5)))  # noqa: S311
+
+
+def human_press_key(page: Page, key: str, count: int = 1, delay: int = 100, sleep=True):
+    for _ in range(count):
+        page.keyboard.press(key, delay=randint(int(delay * 0.5), int(delay * 1.5)))  # noqa: S311
+        if sleep:
+            time.sleep(randint(int(delay * 1.5), int(delay * 2.5)) / 1000)  # noqa: S311