dhisana 0.0.1.dev302__tar.gz → 0.0.1.dev304__tar.gz

{dhisana-0.0.1.dev302 → dhisana-0.0.1.dev304}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dhisana
-Version: 0.0.1.dev302
+Version: 0.0.1.dev304
 Summary: A Python SDK for Dhisana AI Platform
 Home-page: https://github.com/dhisana-ai/dhisana-python-sdk
 Author: Admin

{dhisana-0.0.1.dev302 → dhisana-0.0.1.dev304}/setup.py

@@ -2,7 +2,7 @@ from setuptools import setup, find_packages
 
 setup(
     name='dhisana',  
-    version='0.0.1-dev302',
+    version='0.0.1-dev304',
     description='A Python SDK for Dhisana AI Platform',
     author='Admin',
     author_email='contact@dhisana.ai',

{dhisana-0.0.1.dev302 → dhisana-0.0.1.dev304}/src/dhisana/utils/enrich_lead_information.py

@@ -658,9 +658,57 @@ async def enrich_user_info(
                 tool_config=tool_config,
             )
             if found_linkedin_url:
-                user_linkedin_url = found_linkedin_url
-                input_properties["user_linkedin_url"] = user_linkedin_url
-        if not user_linkedin_url and email:
+                should_use_url = True
+                if use_strict_check and org_domain:
+                    try:
+                        temp_props = {"user_linkedin_url": found_linkedin_url}
+                        enriched = await enrich_with_provider(temp_props, tool_config)
+                        enriched_domain = (enriched.get("primary_domain_of_organization") or "").strip().lower()
+                        # If no domain returned but we have an organization LinkedIn URL,
+                        # look up the org to get its domain
+                        if not enriched_domain:
+                            enriched_org_linkedin = (enriched.get("organization_linkedin_url") or "").strip()
+                            if enriched_org_linkedin:
+                                try:
+                                    org_info = await search_organization_by_linkedin_or_domain(
+                                        linkedin_url=enriched_org_linkedin,
+                                        tool_config=tool_config,
+                                    )
+                                    enriched_domain = (org_info.get("domain") or "").strip().lower()
+                                except Exception:
+                                    logger.debug("Could not look up org domain from LinkedIn URL: %s", enriched_org_linkedin)
+                        input_org_domain = org_domain.strip().lower()
+                        if input_org_domain and enriched_domain and enriched_domain != input_org_domain:
+                            logger.info(
+                                "Skipping Google-found LinkedIn URL %s: enriched domain '%s' does not match input domain '%s'",
+                                found_linkedin_url, enriched_domain, input_org_domain,
+                            )
+                            should_use_url = False
+                        elif input_org_domain and not enriched_domain:
+                            logger.info(
+                                "Skipping Google-found LinkedIn URL %s: could not determine enriched org domain to verify against input domain '%s'",
+                                found_linkedin_url, input_org_domain,
+                            )
+                            should_use_url = False
+                    except Exception:
+                        logger.debug("Could not verify LinkedIn URL domain via enrichment; proceeding with found URL.")
+                if should_use_url:
+                    user_linkedin_url = found_linkedin_url
+                    input_properties["user_linkedin_url"] = user_linkedin_url
+        # Skip email-based LinkedIn search if strict check is on and the email
+        # domain doesn't match the expected org domain (would waste API calls
+        # searching for a person at a different company).
+        skip_email_search = False
+        if use_strict_check and org_domain and email:
+            email_domain = email.split("@")[-1].strip().lower()
+            if email_domain and org_domain.strip().lower() != email_domain:
+                skip_email_search = True
+                logger.info(
+                    "Skipping email-based LinkedIn search: email domain '%s' does not match org domain '%s'",
+                    email_domain, org_domain,
+                )
+
+        if not user_linkedin_url and email and not skip_email_search:
             # If we have an email but no LinkedIn URL yet, try searching by email via Google
             email_lookup_result = await find_user_linkedin_url_by_email_google(
                 email=email,

{dhisana-0.0.1.dev302 → dhisana-0.0.1.dev304}/src/dhisana/utils/generate_custom_message.py

@@ -11,6 +11,7 @@ from dhisana.utils.generate_structured_output_internal import (
 )
 from datetime import datetime
 from pydantic import BaseModel, ConfigDict
+from dhisana.utils.generate_email import _sanitize_lead_for_prompt
 
 # -----------------------------------------------------------------------------
 # Custom Message schema
@@ -140,7 +141,7 @@ async def generate_custom_message_copy(
         Below is the context in 6 main sections. Use it to craft a concise, professional message:
 
         1) Lead Information:
-        {lead_data.dict()}
+        {_sanitize_lead_for_prompt(lead_data)}
 
         2) Sender Information:
         Full Name: {sender_data.sender_full_name or ''}

{dhisana-0.0.1.dev302 → dhisana-0.0.1.dev304}/src/dhisana/utils/generate_email.py

@@ -33,6 +33,14 @@ class EmailCopy(BaseModel):
 # Phone-number pattern used to redact lead phone numbers from the prompt
 _PHONE_RE = re.compile(r"(\+?\d[\d\-.\s()]{7,}\d)")
 
+# Large blobs that should never be forwarded into the email-generation prompt
+_FULL_EXPORT_EXCLUDED_ADDITIONAL = {
+    "pc_person_data",
+    "pc_company_data",
+    "apollo_person_data",
+    "apollo_company_data",
+}
+
 
 def _sanitize_lead_for_prompt(lead: "Lead") -> str:
     """Return a clean JSON string of lead data suitable for an LLM prompt.
@@ -53,6 +61,12 @@ def _sanitize_lead_for_prompt(lead: "Lead") -> str:
     if data.get("research_summary"):
         data["research_summary"] = _PHONE_RE.sub("[redacted]", data["research_summary"])
 
+    # Strip bulky raw-export blobs from additional_properties
+    ap = data.get("additional_properties")
+    if isinstance(ap, dict):
+        for key in _FULL_EXPORT_EXCLUDED_ADDITIONAL:
+            ap.pop(key, None)
+
     return json.dumps(data, indent=2, default=str)
 
 

{dhisana-0.0.1.dev302 → dhisana-0.0.1.dev304}/src/dhisana/utils/generate_email_response.py

@@ -11,6 +11,7 @@ from dhisana.schemas.sales import (
     SenderInfo
 )
 from dhisana.utils.assistant_tool_tag import assistant_tool
+from dhisana.utils.generate_email import _sanitize_lead_for_prompt
 from dhisana.utils.generate_structured_output_internal import (
     get_structured_output_with_assistant_and_vector_store,
     get_structured_output_internal
@@ -283,7 +284,7 @@ async def generate_inbound_email_response_copy(
         if cleaned_context.current_conversation_context.current_email_thread else []}
 
     2) Lead information:
-    {lead_data.model_dump()}
+    {_sanitize_lead_for_prompt(lead_data)}
 
     Sender information:
     - Full name: {sender_data.sender_full_name or ''}

{dhisana-0.0.1.dev302 → dhisana-0.0.1.dev304}/src/dhisana/utils/generate_linkedin_response_message.py

@@ -14,6 +14,7 @@ from dhisana.utils.generate_structured_output_internal import (
     get_structured_output_with_assistant_and_vector_store
 )
 from dhisana.utils.assistant_tool_tag import assistant_tool
+from dhisana.utils.generate_email import _sanitize_lead_for_prompt
 import datetime
 
 # ---------------------------------------------------------------------------------------
@@ -112,7 +113,7 @@ async def generate_linkedin_response_message_copy(
 
         2.
         Lead Information:
-        {lead_data.dict()}
+        {_sanitize_lead_for_prompt(lead_data)}
 
         Sender Information:
         Full Name: {sender_data.sender_full_name or ''}

{dhisana-0.0.1.dev302 → dhisana-0.0.1.dev304}/src/dhisana/utils/serpapi_search_tools.py

@@ -401,12 +401,6 @@ async def find_user_linkedin_url_by_email_google(
         if query and query not in queries:
             queries.append(query)
 
-    def add_query_parts(*parts: str) -> None:
-        tokens = [part.strip() for part in parts if part and part.strip()]
-        if not tokens:
-            return
-        add_query(" ".join(tokens))
-
     enriched_terms = []
     if user_name:
         enriched_terms.append(f'"{user_name}"')
@@ -418,47 +412,23 @@ async def find_user_linkedin_url_by_email_google(
         enriched_terms.append(f'"{user_location}"')
     base_hint = " ".join(enriched_terms)
 
-    # Prioritise the direct email search variants before broader fallbacks.
-    add_query_parts(normalized_email, "linkedin.com/in", base_hint)
-    add_query_parts(normalized_email, "linkedin.com", base_hint)
-    add_query_parts(normalized_email, "linkedin", base_hint)
-    add_query_parts(normalized_email, base_hint)
-    add_query(f'"{normalized_email}" "linkedin.com/in" {base_hint}')
-    add_query(f'"{normalized_email}" "linkedin.com" {base_hint}')
-    add_query(f'"{normalized_email}" linkedin {base_hint}')
+    # 1) Best query: site-scoped with exact email
+    add_query(f'site:linkedin.com/in "{normalized_email}" {base_hint}')
 
-    if email_local_part and email_local_part != normalized_email:
-        add_query_parts(email_local_part, "linkedin.com/in", base_hint)
-        add_query_parts(email_local_part, "linkedin.com", base_hint)
-        add_query_parts(email_local_part, "linkedin", base_hint)
-        add_query(f'"{email_local_part}" "linkedin.com/in" {base_hint}')
-        add_query(f'"{email_local_part}" "linkedin.com" {base_hint}')
+    # 2) Exact email with linkedin scope (Google may surface LinkedIn even without site:)
+    add_query(f'"{normalized_email}" linkedin.com/in {base_hint}')
 
-    if email_local_humanized and email_local_humanized not in {email_local_part, normalized_email}:
-        add_query_parts(email_local_humanized, "linkedin", base_hint)
-        add_query(f'"{email_local_humanized}" linkedin {base_hint}')
+    # 3) Unquoted email with linkedin scope (catches partial matches)
+    add_query(f'{normalized_email} linkedin.com/in {base_hint}')
 
-    if normalized_email:
-        add_query(f'site:linkedin.com/in "{normalized_email}" {base_hint}')
-
-    if email_local_part:
+    # 4) Local part fallback with site scope (e.g. "santosh" on linkedin)
+    if email_local_part and email_local_part != normalized_email:
         add_query(f'site:linkedin.com/in "{email_local_part}" {base_hint}')
 
-    if email_local_humanized and email_local_humanized != email_local_part:
+    # 5) Humanized local part fallback (e.g. "john smith" from "john.smith")
+    if email_local_humanized and email_local_humanized not in {email_local_part, normalized_email}:
         add_query(f'site:linkedin.com/in "{email_local_humanized}" {base_hint}')
 
-    if base_hint:
-        lookup_hint = user_name or email_local_humanized or email_local_part or normalized_email
-        add_query(
-            f'site:linkedin.com/in "{normalized_email}" {base_hint} '
-            f'intitle:"{lookup_hint}" -intitle:"profiles"'
-        )
-        if email_local_humanized:
-            add_query(
-                f'site:linkedin.com/in "{email_local_humanized}" {base_hint} '
-                f'intitle:"{lookup_hint}" -intitle:"profiles"'
-            )
-
     candidate_records: List[Dict[str, str]] = []
     seen_links: Set[str] = set()
     best_llm_choice: Optional[LinkedinCandidateChoice] = None

{dhisana-0.0.1.dev302 → dhisana-0.0.1.dev304}/src/dhisana.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dhisana
-Version: 0.0.1.dev302
+Version: 0.0.1.dev304
 Summary: A Python SDK for Dhisana AI Platform
 Home-page: https://github.com/dhisana-ai/dhisana-python-sdk
 Author: Admin