fraudcrawler 0.6.1__py3-none-any.whl → 0.6.3__py3-none-any.whl

fraudcrawler/__init__.py

@@ -2,7 +2,9 @@ from fraudcrawler.scraping.search import Searcher, SearchEngineName
 from fraudcrawler.scraping.enrich import Enricher
 from fraudcrawler.scraping.url import URLCollector
 from fraudcrawler.scraping.zyte import ZyteAPI
+from fraudcrawler.scraping.config import ScrapingConfig
 from fraudcrawler.processing.processor import Processor
+from fraudcrawler.processing.config import ProcessingConfig
 from fraudcrawler.base.orchestrator import Orchestrator
 from fraudcrawler.base.client import FraudCrawlerClient
 from fraudcrawler.base.base import (
@@ -22,7 +24,9 @@ __all__ = [
     "Enricher",
     "URLCollector",
     "ZyteAPI",
+    "ScrapingConfig",
     "Processor",
+    "ProcessingConfig",
     "Orchestrator",
     "ProductItem",
     "FraudCrawlerClient",

fraudcrawler/base/base.py

@@ -45,6 +45,7 @@ class Setup(BaseSettings):
     dataforseo_pwd: str
     zyteapi_key: str
     openaiapi_key: str
+    pypy_token: str
 
     class Config:
         env_file = ".env"
@@ -140,6 +141,8 @@ class ProductItem(BaseModel):
     url_resolved: str
     search_engine_name: str
     domain: str
+    exact_search: bool = False
+    exact_search_match: bool = False
 
     # Context parameters
     product_name: str | None = None

fraudcrawler/base/client.py

@@ -19,7 +19,9 @@ from fraudcrawler.base.base import (
     ProductItem,
 )
 from fraudcrawler.base.orchestrator import Orchestrator
+from fraudcrawler.scraping.config import ScrapingConfig
 from fraudcrawler.scraping.search import SearchEngineName
+from fraudcrawler.processing.config import ProcessingConfig
 
 logger = logging.getLogger(__name__)
 
@@ -141,15 +143,19 @@ class FraudCrawlerClient(Orchestrator):
 
         asyncio.run(
             _run(
-                search_term=search_term,
-                search_engines=nrm_search_engines,
-                language=language,
-                location=location,
-                deepness=deepness,
-                prompts=prompts,
-                marketplaces=marketplaces,
-                excluded_urls=excluded_urls,
-                previously_collected_urls=previously_collected_urls,
+                scraping_config=ScrapingConfig(
+                    search_term=search_term,
+                    search_engines=nrm_search_engines,
+                    language=language,
+                    location=location,
+                    deepness=deepness,
+                    marketplaces=marketplaces,
+                    excluded_urls=excluded_urls,
+                    previously_collected_urls=previously_collected_urls,
+                ),
+                processing_config=ProcessingConfig(
+                    prompts=prompts,
+                ),
             )
         )
 

fraudcrawler/base/orchestrator.py

@@ -3,10 +3,12 @@ import asyncio
 import logging
 from typing import cast, Dict, List, Self
 
-from bs4 import BeautifulSoup
 import httpx
+import re
 
 from fraudcrawler.settings import (
+    EXACT_MATCH_PRODUCT_FIELDS,
+    EXACT_MATCH_FIELD_SEPARATOR,
     PROCESSOR_DEFAULT_MODEL,
 )
 from fraudcrawler.settings import (
@@ -15,11 +17,9 @@ from fraudcrawler.settings import (
     DEFAULT_N_PROC_WKRS,
 )
 from fraudcrawler.base.base import (
-    Deepness,
     Host,
     Language,
     Location,
-    Prompt,
     ProductItem,
     HttpxAsyncClient,
 )
@@ -27,9 +27,11 @@ from fraudcrawler import (
     Searcher,
     SearchEngineName,
     Enricher,
-    URLCollector,
     ZyteAPI,
+    URLCollector,
+    ScrapingConfig,
     Processor,
+    ProcessingConfig,
 )
 
 logger = logging.getLogger(__name__)
@@ -227,44 +229,29 @@ class Orchestrator(ABC):
 
             if not product.filtered:
                 try:
-                    # Fetch the product context from Zyte API
+                    # Fetch and enrich the product context from Zyte API
                     details = await self._zyteapi.details(url=product.url)
-                    url_resolved = self._zyteapi.extract_url_resolved(details=details)
-                    if url_resolved:
-                        product.url_resolved = url_resolved
-                    product.product_name = self._zyteapi.extract_product_name(
-                        details=details
+                    product = self._zyteapi.enrich_context(
+                        product=product, details=details
                     )
 
-                    # If the resolved URL is different from the original URL, we also need to update the domain as
-                    # otherwise the unresolved domain will be shown.
-                    # For example for an unresolved domain "toppreise.ch" but resolved "digitec.ch
-                    if url_resolved and url_resolved != product.url:
-                        logger.debug(
-                            f"URL resolved for {product.url} is {url_resolved}"
-                        )
-                        product.domain = self._searcher._get_domain(url_resolved)
-
-                    product.product_price = self._zyteapi.extract_product_price(
-                        details=details
-                    )
-                    product.product_description = (
-                        self._zyteapi.extract_product_description(details=details)
-                    )
-                    product.product_images = self._zyteapi.extract_image_urls(
-                        details=details
-                    )
-                    product.probability = self._zyteapi.extract_probability(
-                        details=details
-                    )
-                    product.html = self._zyteapi.extract_html(details=details)
-                    if product.html:
-                        soup = BeautifulSoup(product.html, "html.parser")
-                        product.html_clean = soup.get_text(separator=" ", strip=True)
                     # Filter the product based on the probability threshold
                     if not self._zyteapi.keep_product(details=details):
                         product.filtered = True
-                        product.filtered_at_stage = "Zyte probability threshold"
+                        product.filtered_at_stage = (
+                            "Context (Zyte probability threshold)"
+                        )
+
+                    # Check for exact match inside the full product context
+                    product = self._check_exact_search(product=product)
+                    if (
+                        not product.filtered
+                        and product.exact_search
+                        and not product.exact_search_match
+                    ):
+                        product.filtered = True
+                        product.filtered_at_stage = "Context (exact search)"
+
                 except Exception as e:
                     logger.warning(f"Error executing Zyte API search: {e}.")
             await queue_out.put(product)
@@ -274,14 +261,14 @@ class Orchestrator(ABC):
         self,
         queue_in: asyncio.Queue[ProductItem | None],
         queue_out: asyncio.Queue[ProductItem | None],
-        prompts: List[Prompt],
+        processing_config: ProcessingConfig,
     ) -> None:
         """Collects the product details from the queue_in, processes them (filtering, relevance, etc.) and puts the results into queue_out.
 
         Args:
             queue_in: The input queue containing the product details.
             queue_out: The output queue to put the processed product details.
-            prompts: The list of prompts to use for classification.
+            processing_config: Sets up the processing pipeline step.
         """
 
         # Process the products
@@ -295,7 +282,7 @@ class Orchestrator(ABC):
             if not product.filtered:
                 try:
                     # Run all the configured prompts
-                    for prompt in prompts:
+                    for prompt in processing_config.prompts:
                         classification = await self._processor.classify(
                             product=product,
                             prompt=prompt,
@@ -331,7 +318,7 @@ class Orchestrator(ABC):
         n_srch_wkrs: int,
         n_cntx_wkrs: int,
         n_proc_wkrs: int,
-        prompts: List[Prompt],
+        processing_config: ProcessingConfig,
     ) -> None:
         """Sets up the necessary queues and workers for the async framework.
 
@@ -339,7 +326,7 @@ class Orchestrator(ABC):
             n_srch_wkrs: Number of async workers for search.
             n_cntx_wkrs: Number of async workers for context extraction.
             n_proc_wkrs: Number of async workers for processing.
-            prompts: The list of prompts used for the classification by func:`Processor.classify`.
+            processing_config: Sets up the processing pipeline step.
         """
 
         # Setup the input/output queues for the workers
@@ -382,7 +369,7 @@ class Orchestrator(ABC):
                 self._proc_execute(
                     queue_in=proc_queue,
                     queue_out=res_queue,
-                    prompts=prompts,
+                    processing_config=processing_config,
                 )
             )
             for _ in range(n_proc_wkrs)
@@ -436,13 +423,7 @@ class Orchestrator(ABC):
     async def _add_srch_items(
         self,
         queue: asyncio.Queue[dict | None],
-        search_term: str,
-        search_engines: List[SearchEngineName],
-        language: Language,
-        location: Location,
-        deepness: Deepness,
-        marketplaces: List[Host] | None,
-        excluded_urls: List[Host] | None,
+        scraping_config: ScrapingConfig,
     ) -> None:
         """Adds all the (enriched) search_term (as srch items) to the queue.
 
@@ -461,12 +442,17 @@ class Orchestrator(ABC):
                 for each search_engine
                     add item to queue
         """
+        search_term = scraping_config.search_term
+        search_engines = scraping_config.search_engines
+        language = scraping_config.language
+        location = scraping_config.location
+        deepness = scraping_config.deepness
         common_kwargs = {
             "queue": queue,
             "language": language,
             "location": location,
-            "marketplaces": marketplaces,
-            "excluded_urls": excluded_urls,
+            "marketplaces": scraping_config.marketplaces,
+            "excluded_urls": scraping_config.excluded_urls,
         }
 
         # Add initial items to the queue
@@ -502,48 +488,101 @@ class Orchestrator(ABC):
                         **common_kwargs,  # type: ignore[arg-type]
                     )
 
+    @staticmethod
+    def _is_exact_search(search_term: str) -> bool:
+        """Check if the search term is an exact search (contains double quotation marks).
+
+        Args:
+            search_term: The search term to check.
+        """
+        return '"' in search_term
+
+    @staticmethod
+    def _extract_exact_search_terms(search_term: str) -> list[str]:
+        """Extract all exact search terms from within double quotation marks (empty if no quotes found).
+
+        Args:
+            search_term: The search term that may contain double quotation marks.
+        """
+        # Find all double-quoted strings
+        double_quote_matches = re.findall(r'"([^"]*)"', search_term)
+        return double_quote_matches
+
+    @staticmethod
+    def _check_exact_search_terms_match(
+        product: ProductItem,
+        exact_search_terms: list[str],
+    ) -> bool:
+        """Check if the product, represented by a string of selected attributes, matches ALL of the exact search terms.
+
+        Args:
+            product: The product item.
+            exact_search_terms: List of exact search terms to match against.
+        """
+        field_values = [
+            str(val)
+            for fld in EXACT_MATCH_PRODUCT_FIELDS
+            if (val := getattr(product, fld, None)) is not None
+        ]
+        product_str_lower = EXACT_MATCH_FIELD_SEPARATOR.join(field_values).lower()
+
+        return all(
+            re.search(re.escape(est.lower()), product_str_lower)
+            for est in exact_search_terms
+        )
+
+    def _check_exact_search(self, product: ProductItem) -> ProductItem:
+        """Checks if the search term requests an exact search and if yes, checks for conformity."""
+        # Check for exact search and apply regex matching
+        exact_search = self._is_exact_search(product.search_term)
+        product.exact_search = exact_search
+
+        # Only set exact_search_match if this was an exact search (contains quotes)
+        if exact_search:
+            exact_search_terms = self._extract_exact_search_terms(product.search_term)
+            if exact_search_terms:
+                product.exact_search_match = self._check_exact_search_terms_match(
+                    product=product, exact_search_terms=exact_search_terms
+                )
+                logger.debug(
+                    f"Exact search terms {exact_search_terms} matched: {product.exact_search_match} "
+                    f"for offer with url={product.url}"
+                )
+            else:
+                logger.warning(
+                    f"is_exact_search=True but no exact search terms found in search_term='{product.search_term}' "
+                    f"for offer with url={product.url}"
+                )
+        # If exact_search is False, product.exact_search_match remains False (default value)
+        return product
+
     async def run(
         self,
-        search_term: str,
-        search_engines: List[SearchEngineName],
-        language: Language,
-        location: Location,
-        deepness: Deepness,
-        prompts: List[Prompt],
-        marketplaces: List[Host] | None = None,
-        excluded_urls: List[Host] | None = None,
-        previously_collected_urls: List[str] | None = None,
+        scraping_config: ScrapingConfig,
+        processing_config: ProcessingConfig,
     ) -> None:
         """Runs the pipeline steps: srch, deduplication, context extraction, processing, and collect the results.
 
         Args:
-            search_term: The search term for the query.
-            search_engines: The list of search engines to use for the search query.
-            language: The language to use for the query.
-            location: The location to use for the query.
-            deepness: The search depth and enrichment details.
-            prompts: The list of prompt to use for classification.
-            marketplaces: The marketplaces to include in the search.
-            excluded_urls: The URLs to exclude from the search.
-            previously_collected_urls: The urls that have been collected previously and are ignored.
+            scraping_config: Sets up the scraping pipeline step.
+            processing_config: Sets up the processing pipeline step.
         """
         # ---------------------------
         #        INITIAL SETUP
         # ---------------------------
-        # Ensure we have at least one search engine
-        if not search_engines:
+        # Ensure we have at least one search engine (the list might be empty)
+        if not scraping_config.search_engines:
             logger.warning(
                 "No search engines specified, using all available search engines"
             )
-            search_engines = list(SearchEngineName)
+            scraping_config.search_engines = list(SearchEngineName)
 
         # Handle previously collected URLs
-        if previously_collected_urls:
-            self._url_collector.add_previously_collected_urls(
-                urls=previously_collected_urls
-            )
+        if pcurls := scraping_config.previously_collected_urls:
+            self._url_collector.add_previously_collected_urls(urls=pcurls)
 
         # Setup the async framework
+        deepness = scraping_config.deepness
         n_terms_max = 1 + (
             deepness.enrichment.additional_terms if deepness.enrichment else 0
         )
@@ -558,7 +597,7 @@ class Orchestrator(ABC):
             n_srch_wkrs=n_srch_wkrs,
             n_cntx_wkrs=n_cntx_wkrs,
             n_proc_wkrs=n_proc_wkrs,
-            prompts=prompts,
+            processing_config=processing_config,
         )
 
         # Check setup of async framework
@@ -581,13 +620,7 @@ class Orchestrator(ABC):
         srch_queue = self._queues["srch"]
         await self._add_srch_items(
             queue=srch_queue,
-            search_term=search_term,
-            search_engines=search_engines,
-            language=language,
-            location=location,
-            deepness=deepness,
-            marketplaces=marketplaces,
-            excluded_urls=excluded_urls,
+            scraping_config=scraping_config,
         )
 
         # -----------------------------

fraudcrawler/launch_demo_pipeline.py

@@ -97,4 +97,4 @@ def search(search_term: str):
 
 
 if __name__ == "__main__":
-    search(search_term='Gorenje "R619FEW5"')
+    search(search_term="Kaffeebohnen")

fraudcrawler/processing/config.py

@@ -0,0 +1,12 @@
+from pydantic import BaseModel, Field
+from typing import List
+
+from fraudcrawler.base.base import Prompt
+
+
+class ProcessingConfig(BaseModel):
+    """Sets up the processing pipeline step."""
+
+    prompts: List[Prompt] = Field(
+        description="The list of prompts to use for classification."
+    )

fraudcrawler/scraping/config.py

@@ -0,0 +1,32 @@
+from pydantic import BaseModel, Field
+from typing import List
+
+from fraudcrawler.scraping.search import SearchEngineName
+from fraudcrawler.base.base import (
+    Language,
+    Location,
+    Deepness,
+    Host,
+)
+
+
+class ScrapingConfig(BaseModel):
+    """Sets up the scraping pipeline step."""
+
+    search_term: str = Field(description="The search term for the query.")
+    search_engines: List[SearchEngineName] = Field(
+        description="The list of search engines to use for the search query."
+    )
+    language: Language = Field(description="The language to use for the query.")
+    location: Location = Field(description="The location to use for the query.")
+    deepness: Deepness = Field(description="The search depth and enrichment details.")
+    marketplaces: List[Host] | None = Field(
+        default=None, description="The marketplaces to include in the search."
+    )
+    excluded_urls: List[Host] | None = Field(
+        default=None, description="The URLs to exclude from the search."
+    )
+    previously_collected_urls: List[str] | None = Field(
+        default=None,
+        description="The URLs that have been collected previously and are ignored.",
+    )

fraudcrawler/scraping/zyte.py

@@ -2,11 +2,12 @@ from base64 import b64decode
 import logging
 from typing import List
 
+from bs4 import BeautifulSoup
 import httpx
 from tenacity import RetryCallState
 
 from fraudcrawler.settings import ZYTE_DEFALUT_PROBABILITY_THRESHOLD
-from fraudcrawler.base.base import DomainUtils
+from fraudcrawler.base.base import DomainUtils, ProductItem
 from fraudcrawler.base.retry import get_async_retry
 
 logger = logging.getLogger(__name__)
@@ -61,77 +62,8 @@ class ZyteAPI(DomainUtils):
         else:
             logger.debug(f"retry_state is {retry_state}; not logging before_sleep.")
 
-    async def details(self, url: str) -> dict:
-        """Fetches product details for a single URL.
-
-        Args:
-            url: The URL to fetch product details from.
-
-        Returns:
-            A dictionary containing the product details, fields include:
-            (c.f. https://docs.zyte.com/zyte-api/usage/reference.html#operation/extract/response/200/product)
-            {
-                "url": str,
-                "statusCode": str,
-                "product": {
-                    "name": str,
-                    "price": str,
-                    "mainImage": {"url": str},
-                    "images": [{"url": str}],
-                    "description": str,
-                    "metadata": {
-                        "probability": float,
-                    },
-                },
-                "httpResponseBody": base64
-            }
-        """
-        logger.info(f"Fetching product details by Zyte for URL {url}.")
-
-        # Perform the request and retry if necessary. There is some context aware logging:
-        #  - `before`: before the request is made (and before retrying)
-        #  - `before_sleep`: if the request fails before sleeping
-        retry = get_async_retry()
-        retry.before = lambda retry_state: self._log_before(
-            url=url, retry_state=retry_state
-        )
-        retry.before_sleep = lambda retry_state: self._log_before_sleep(
-            url=url, retry_state=retry_state
-        )
-        async for attempt in retry:
-            with attempt:
-                response = await self._http_client.post(
-                    url=self._endpoint,
-                    json={"url": url, **self._config},
-                    auth=(self._api_key, ""),  # API key as username, empty password
-                )
-                response.raise_for_status()
-
-        details = response.json()
-        return details
-
-    @staticmethod
-    def keep_product(
-        details: dict,
-        threshold: float = ZYTE_DEFALUT_PROBABILITY_THRESHOLD,
-    ) -> bool:
-        """Determines whether to keep the product based on the probability threshold.
-
-        Args:
-            details: A product details data dictionary.
-            threshold: The probability threshold used to filter the products.
-        """
-        try:
-            prob = float(details["product"]["metadata"]["probability"])
-        except KeyError:
-            logger.warning(
-                f"Product with url={details.get('url')} has no probability value - product is ignored"
-            )
-            return False
-        return prob > threshold
-
     @staticmethod
-    def extract_product_name(details: dict) -> str | None:
+    def _extract_product_name(details: dict) -> str | None:
         """Extracts the product name from the product data.
 
         The input argument is a dictionary of the following structure:
@@ -144,7 +76,7 @@ class ZyteAPI(DomainUtils):
         return details.get("product", {}).get("name")
 
     @staticmethod
-    def extract_url_resolved(details: dict) -> str | None:
+    def _extract_url_resolved(details: dict) -> str | None:
         """Extracts the resolved URL from the product data - this is automatically resolved by Zyte.
 
         The input argument is a dictionary of the following structure:
@@ -157,7 +89,7 @@ class ZyteAPI(DomainUtils):
         return details.get("product", {}).get("url")
 
     @staticmethod
-    def extract_product_price(details: dict) -> str | None:
+    def _extract_product_price(details: dict) -> str | None:
         """Extracts the product price from the product data.
 
         The input argument is a dictionary of the following structure:
@@ -170,7 +102,7 @@ class ZyteAPI(DomainUtils):
         return details.get("product", {}).get("price")
 
     @staticmethod
-    def extract_product_description(details: dict) -> str | None:
+    def _extract_product_description(details: dict) -> str | None:
         """Extracts the product description from the product data.
 
         The input argument is a dictionary of the following structure:
@@ -183,7 +115,7 @@ class ZyteAPI(DomainUtils):
         return details.get("product", {}).get("description")
 
     @staticmethod
-    def extract_image_urls(details: dict) -> List[str]:
+    def _extract_image_urls(details: dict) -> List[str]:
         """Extracts the images from the product data.
 
         The input argument is a dictionary of the following structure:
@@ -206,7 +138,7 @@ class ZyteAPI(DomainUtils):
         return images
 
     @staticmethod
-    def extract_probability(details: dict) -> float:
+    def _extract_probability(details: dict) -> float:
         """Extracts the probability from the product data.
 
         The input argument is a dictionary of the following structure:
@@ -223,7 +155,7 @@ class ZyteAPI(DomainUtils):
         )
 
     @staticmethod
-    def extract_html(details: dict) -> str | None:
+    def _extract_html(details: dict) -> str | None:
         """Extracts the HTML from the Zyte API response.
 
         The input argument is a dictionary of the following structure:
@@ -243,6 +175,51 @@ class ZyteAPI(DomainUtils):
             return decoded_string
         return None
 
+    def enrich_context(self, product: ProductItem, details: dict) -> ProductItem:
+        product.product_name = self._extract_product_name(details=details)
+
+        url_resolved = self._extract_url_resolved(details=details)
+        if url_resolved:
+            product.url_resolved = url_resolved
+
+        # If the resolved URL is different from the original URL, we also need to update the domain as
+        # otherwise the unresolved domain will be shown.
+        # For example for an unresolved domain "toppreise.ch" but resolved "digitec.ch
+        if url_resolved and url_resolved != product.url:
+            logger.debug(f"URL resolved for {product.url} is {url_resolved}")
+            product.domain = self._get_domain(url=url_resolved)
+
+        product.product_price = self._extract_product_price(details=details)
+        product.product_description = self._extract_product_description(details=details)
+        product.product_images = self._extract_image_urls(details=details)
+        product.probability = self._extract_probability(details=details)
+        product.html = self._extract_html(details=details)
+        if product.html:
+            soup = BeautifulSoup(product.html, "html.parser")
+            product.html_clean = soup.get_text(separator=" ", strip=True)
+
+        return product
+
+    @staticmethod
+    def keep_product(
+        details: dict,
+        threshold: float = ZYTE_DEFALUT_PROBABILITY_THRESHOLD,
+    ) -> bool:
+        """Determines whether to keep the product based on the probability threshold.
+
+        Args:
+            details: A product details data dictionary.
+            threshold: The probability threshold used to filter the products.
+        """
+        try:
+            prob = float(details["product"]["metadata"]["probability"])
+        except KeyError:
+            logger.warning(
+                f"Product with url={details.get('url')} has no probability value - product is ignored"
+            )
+            return False
+        return prob > threshold
+
     async def unblock_url_content(self, url: str) -> bytes:
         """Unblock the content of an URL using Zyte proxy mode.
 
@@ -256,3 +233,52 @@ class ZyteAPI(DomainUtils):
             raise httpx.HTTPError("No httpResponseBody in Zyte response")
 
         return b64decode(details["httpResponseBody"])
+
+    async def details(self, url: str) -> dict:
+        """Fetches product details for a single URL.
+
+        Args:
+            url: The URL to fetch product details from.
+
+        Returns:
+            A dictionary containing the product details, fields include:
+            (c.f. https://docs.zyte.com/zyte-api/usage/reference.html#operation/extract/response/200/product)
+            {
+                "url": str,
+                "statusCode": str,
+                "product": {
+                    "name": str,
+                    "price": str,
+                    "mainImage": {"url": str},
+                    "images": [{"url": str}],
+                    "description": str,
+                    "metadata": {
+                        "probability": float,
+                    },
+                },
+                "httpResponseBody": base64
+            }
+        """
+        logger.info(f"Fetching product details by Zyte for URL {url}.")
+
+        # Perform the request and retry if necessary. There is some context aware logging:
+        #  - `before`: before the request is made (and before retrying)
+        #  - `before_sleep`: if the request fails before sleeping
+        retry = get_async_retry()
+        retry.before = lambda retry_state: self._log_before(
+            url=url, retry_state=retry_state
+        )
+        retry.before_sleep = lambda retry_state: self._log_before_sleep(
+            url=url, retry_state=retry_state
+        )
+        async for attempt in retry:
+            with attempt:
+                response = await self._http_client.post(
+                    url=self._endpoint,
+                    json={"url": url, **self._config},
+                    auth=(self._api_key, ""),  # API key as username, empty password
+                )
+                response.raise_for_status()
+
+        details = response.json()
+        return details

fraudcrawler/settings.py

@@ -78,6 +78,14 @@ ENRICHMENT_DEFAULT_LIMIT = 10
 # Zyte settings
 ZYTE_DEFALUT_PROBABILITY_THRESHOLD = 0.1
 
+# Exact match settings
+EXACT_MATCH_PRODUCT_FIELDS = {
+    "url_resolvedproduct_name",
+    "product_description",
+    "html",
+}
+EXACT_MATCH_FIELD_SEPARATOR = "\n"
+
 # Processor settings
 PROCESSOR_DEFAULT_MODEL = "gpt-4o"
 PROCESSOR_DEFAULT_IF_MISSING = -1

{fraudcrawler-0.6.1.dist-info → fraudcrawler-0.6.3.dist-info}/METADATA

@@ -1,9 +1,9 @@
-Metadata-Version: 2.3
+Metadata-Version: 2.4
 Name: fraudcrawler
-Version: 0.6.1
+Version: 0.6.3
 Summary: Intelligent Market Monitoring
-Home-page: https://github.com/open-veanu/fraudcrawler
 License: MIT
+License-File: LICENSE
 Author: Domingo Bertus
 Author-email: hello@veanu.ch
 Requires-Python: >=3.11,<4.0
@@ -12,6 +12,7 @@ Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
 Requires-Dist: beautifulsoup4 (>=4.13.4,<5.0.0)
 Requires-Dist: httpx (>=0.28.1,<0.29.0)
 Requires-Dist: openai (>=1.68.2,<2.0.0)

fraudcrawler-0.6.3.dist-info/RECORD

@@ -0,0 +1,24 @@
+fraudcrawler/__init__.py,sha256=YEbaofjs8pKkwqz4T-kGk7vHIQ_3XtDlF6D63wfuXjE,1008
+fraudcrawler/base/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+fraudcrawler/base/base.py,sha256=l3rsXKYKmN_I0GkAXDjpjh_s07cE4siFAwMq0byroQM,7815
+fraudcrawler/base/client.py,sha256=6xAhQ7hdWYa7CQ84Ps1XUompCTXiSk0e3PywmTMUGng,6146
+fraudcrawler/base/google-languages.json,sha256=z0VtjMCsCcZq11OkCIb9jMDD1p9Ty4lhV7bq4ddYvec,10748
+fraudcrawler/base/google-locations.json,sha256=UtNu0iSStllvFRTQXMobWKmZR1hKmtgtHftLNgaJTT0,9204345
+fraudcrawler/base/orchestrator.py,sha256=UxjJMLm2kpxG76m2TXShuLkqhUxSuYvZyPP2yy708JA,28082
+fraudcrawler/base/retry.py,sha256=1Ox7RsnnF62dP53rkidRHetA5mr2HS1R-7FskCVbwug,1178
+fraudcrawler/launch_demo_pipeline.py,sha256=_aDqaPdxE_DMwQY5_vpqF2YjwLkWIZq5Z9Tz3sqLKdg,4629
+fraudcrawler/processing/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+fraudcrawler/processing/config.py,sha256=xqqTXK7zFA-7zwk76eZwrF97NtMzMOipUY6imeBIjQ8,301
+fraudcrawler/processing/processor.py,sha256=zetp_G5g4z8sBUq-5qOxVRF2W2h9FIwolVxvMqhTmXs,7619
+fraudcrawler/scraping/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+fraudcrawler/scraping/config.py,sha256=8gyfB0VLi_FZr4J7a-HCTSYt8bRgSKXHo-Y9tlsD2MQ,1179
+fraudcrawler/scraping/enrich.py,sha256=M4ErLF7q-5EKfEd-dIDS73mZc-aPFp5pJbgdRKCO3f8,13112
+fraudcrawler/scraping/search.py,sha256=Anm8ymjCH3BVttogHY-_03YRc64yJswJ8OP8DW56O48,34546
+fraudcrawler/scraping/url.py,sha256=unUoZ-bThU99ZlLdDUILdPx1kbtwMWPZVPCDqPscqHw,3217
+fraudcrawler/scraping/zyte.py,sha256=sYpfwMuGE9MYpKvma_8x5Th2VBFn25Mqb4Wd7UChL_g,10215
+fraudcrawler/settings.py,sha256=9ukAkxEzDtvy3xA-jSF3asr9uLIAATNQ-FqrsgCEDUk,4038
+fraudcrawler-0.6.3.dist-info/METADATA,sha256=pP2S9-MFvCkNwWO7YB9Q9oYmZhsHlIxyljFXvTIfgus,6723
+fraudcrawler-0.6.3.dist-info/WHEEL,sha256=zp0Cn7JsFoX2ATtOhtaFYIiE2rmFAD4OcMhtUki8W3U,88
+fraudcrawler-0.6.3.dist-info/entry_points.txt,sha256=1Befm7cM6945y2AA1z9V4gZV63mtSWcAs7ypvgux_Xg,79
+fraudcrawler-0.6.3.dist-info/licenses/LICENSE,sha256=B-3FuHfe3S0fWAlKlceskPcRhzXq81g-rJ-ddUYb4O8,1062
+fraudcrawler-0.6.3.dist-info/RECORD,,

{fraudcrawler-0.6.1.dist-info → fraudcrawler-0.6.3.dist-info}/WHEEL

@@ -1,4 +1,4 @@
 Wheel-Version: 1.0
-Generator: poetry-core 2.0.0
+Generator: poetry-core 2.2.1
 Root-Is-Purelib: true
 Tag: py3-none-any

fraudcrawler-0.6.1.dist-info/RECORD

@@ -1,22 +0,0 @@
-fraudcrawler/__init__.py,sha256=oSwuiyVBBk_HZfeZxXJR0ELtA4mc-upsBMVHSwuokEo,846
-fraudcrawler/base/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-fraudcrawler/base/base.py,sha256=74qwevU8sZBvXAladam0rmjcdn3AiT39MScpxZtD95I,7727
-fraudcrawler/base/client.py,sha256=obxrd65pYja--XQbgpIMsMO6erMNdRG68SzNUs_YvLM,5856
-fraudcrawler/base/google-languages.json,sha256=z0VtjMCsCcZq11OkCIb9jMDD1p9Ty4lhV7bq4ddYvec,10748
-fraudcrawler/base/google-locations.json,sha256=UtNu0iSStllvFRTQXMobWKmZR1hKmtgtHftLNgaJTT0,9204345
-fraudcrawler/base/orchestrator.py,sha256=n0xrMJ9a3g3cRAMmhKEgyrwwrbgsaMno9DeyE93jB5U,27006
-fraudcrawler/base/retry.py,sha256=1Ox7RsnnF62dP53rkidRHetA5mr2HS1R-7FskCVbwug,1178
-fraudcrawler/launch_demo_pipeline.py,sha256=TqlQrs8raT9jIJ3TJK3BOQMLm2qNn2dKaMGL-MyhC70,4635
-fraudcrawler/processing/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-fraudcrawler/processing/processor.py,sha256=zetp_G5g4z8sBUq-5qOxVRF2W2h9FIwolVxvMqhTmXs,7619
-fraudcrawler/scraping/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-fraudcrawler/scraping/enrich.py,sha256=M4ErLF7q-5EKfEd-dIDS73mZc-aPFp5pJbgdRKCO3f8,13112
-fraudcrawler/scraping/search.py,sha256=Anm8ymjCH3BVttogHY-_03YRc64yJswJ8OP8DW56O48,34546
-fraudcrawler/scraping/url.py,sha256=unUoZ-bThU99ZlLdDUILdPx1kbtwMWPZVPCDqPscqHw,3217
-fraudcrawler/scraping/zyte.py,sha256=SxucVH_wtVhPNImIXvijM528IwL6zl6I3ndf0OdVXY0,8860
-fraudcrawler/settings.py,sha256=Bp9_9w_RRr_-PtZXcy30EKbT9YiOc8OLjEMaNZh06vc,3875
-fraudcrawler-0.6.1.dist-info/LICENSE,sha256=B-3FuHfe3S0fWAlKlceskPcRhzXq81g-rJ-ddUYb4O8,1062
-fraudcrawler-0.6.1.dist-info/METADATA,sha256=_LcfOKayMQjAXoCxlJfqYtiSfitegUuQgFUD5XEGFog,6704
-fraudcrawler-0.6.1.dist-info/WHEEL,sha256=RaoafKOydTQ7I_I3JTrPCg6kUmTgtm4BornzOqyEfJ8,88
-fraudcrawler-0.6.1.dist-info/entry_points.txt,sha256=1Befm7cM6945y2AA1z9V4gZV63mtSWcAs7ypvgux_Xg,79
-fraudcrawler-0.6.1.dist-info/RECORD,,