From 20c49d8f728d17e84903c7260ad5b62fb19a26a3 Mon Sep 17 00:00:00 2001
From: Madhu <madhushantangot@gmail.com>
Date: Sat, 11 Jan 2025 21:57:13 +0530
Subject: [PATCH] firecrawl agent implementation

---
 .../__pycache__/firecrawl.cpython-312.pyc     | Bin 0 -> 1618 bytes
 .../ai_competitors_analysis_team/exaai.py     |   2 -
 .../ai_competitors_analysis_team/fire.py      |  41 ++++++++++++++++++
 .../ai_competitors_analysis_team/firecrawl.py |   0
 .../ai_competitors_analysis_team/main.py      |  37 ++++++++++++++++
 .../requirements.txt                          |   3 +-
 6 files changed, 80 insertions(+), 3 deletions(-)
 create mode 100644 ai_agent_tutorials/ai_competitors_analysis_team/__pycache__/firecrawl.cpython-312.pyc
 create mode 100644 ai_agent_tutorials/ai_competitors_analysis_team/fire.py
 delete mode 100644 ai_agent_tutorials/ai_competitors_analysis_team/firecrawl.py

diff --git a/ai_agent_tutorials/ai_competitors_analysis_team/__pycache__/firecrawl.cpython-312.pyc b/ai_agent_tutorials/ai_competitors_analysis_team/__pycache__/firecrawl.cpython-312.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..ccd244c2c6bda1cacd9ba1222f407f6c6d3d42b1
GIT binary patch
literal 1618
zcmah}&1)M+6ra^<Woh+gTXF0-CEehD#0F_`Vz&_5LU58&NbEvWXmnY|J0of0{bFW>
zV&Q;jdhnr_K*52KdrERBJ@#MdMQKu^EEo#yB{v1<R4DY#`b#Z^4$SPEH}CWJ-rJvx
zMH9ig+xWWqV+Nt$)#Y%(Jz0A{!Dj~%L^KciT3yo=o$^w?Ue^;^_tL&mH#C$&R3~X-
zkPI1nl6{=sH&D+2UxRIs+!OP$w$G)RrUr7EC;20}F;bw}Mv4@_N!N4KB&CgW2=Ozd
zT+f4cj^+<_>jk(M$vDkFKo3)g+{Ay87LM<m6PpwBldqJ+c`Y_BHEAG$C|zZox*T`B
z8)4`P;FxO>4K97Q(!#Pk*Kn)n8y{d7FS!>eahKLFxR`vnPL|2?5?&;Wi}?K28?Ovy
zG#!uOh&dZ{GtLOG7}3~B49AAYd?vg3W+baiVbx~=13xK7K*#VyLcQ3K%6nWEt+3-L
z+HtW5jI59O1_?VsTyjG{qLRswJ8kZ{2hb$|D*rm{z3<@IK{8>g<Cfz7YY}Y=jY4%t
z+tO|wlxI}*PU?T`q_$EI5YZp%halh4xAgzA^aX&k+C8zNP_}SDEYbEMFV>}2_pd9r
zCC9F`WQ$sj(DT9$3z`-S8X@<wWMN=6c<3k337V8x1WSC;v<_C)`kaTP4JIPaT&1jB
zG$@vBPQ{4W4eGXn&<mTJmXO%p7<Dotb{!B0oWYJyg}_aq>?jmW#RBw*<Vst}(5Kvz
zR7mE-Gxmh-VnM+{n=tTxB%LG=v0wg2?A_dvz`PzCST$v=^L7wtA@(Q?1f5UwaVQJa
zU%}@bvg5)58jeB(8b@+~05T?y!cM4yDuDjHYkw&y7q*W{t1VhMkT|d}Q?Vh#$c~^(
zxv(&ITkN(CkZ;(yLj@#N^*q0dqexU0ml7K@CkfAyZD_$HQ(zr8rEqY7y-mS{BPsUn
z#^GGmqRn`k?pJw4J_(bbf-t9Gi?875p+BZ(cW&Oh*+0AdL(pHo(trEvv#C#etAonf
z7nOylm4)85HyJej)(*SJhN-b)ZdgL*$?eh*Y16r3Ih!l&>E?8%w>F$Xh0`zcb5HYg
z{q)=(*hX>IaeT}I$LY?rq>RKx+wOEaH6E_B0J9ooVs{HGM?-eeTAeS&xv0fzN#0zI
zQe08@WL9fO337#XJg#1nVXlSn1Hp`|sf^1<V#!zwKB>4mlq<<yMHOJ5I?jHT%g2Bw
z0L2D0bs)b-d=~Whj8e{i1a8_M{e-^@SM?td-@-A}HBH+^r*~0p7rp;FgDPhSg~dUB
zd|;LaX8ujisHA%3*W)NxdTE}xUwm%9H^|TS&C`PuGlSv;yysFqV_2Hh-g&gLhu||@
zK<UDE_ZR)tZ&P+}6--Xo9tF=%e%xCF+Az1T-9OWxIoCJmpQjgIrgPi&uj$!g1*K<`
GX#WO#*T*ve

literal 0
HcmV?d00001

diff --git a/ai_agent_tutorials/ai_competitors_analysis_team/exaai.py b/ai_agent_tutorials/ai_competitors_analysis_team/exaai.py
index 99e2cf3..8a7f630 100644
--- a/ai_agent_tutorials/ai_competitors_analysis_team/exaai.py
+++ b/ai_agent_tutorials/ai_competitors_analysis_team/exaai.py
@@ -1,6 +1,5 @@
 from exa_py import Exa
 
-# Initialize Exa with your API key
 exa = Exa(api_key="EXA_API_KEY")
 
 def get_competitor_urls(url=None, description=None):
@@ -27,6 +26,5 @@ def get_competitor_urls(url=None, description=None):
     # Extracting and return only the competitor URLs
     competitor_urls = [item.url for item in result.results]
     return competitor_urls
-    # return result
 
 print(get_competitor_urls(description="Competitors of a company that works on AI Neuro Health Tech"))
\ No newline at end of file
diff --git a/ai_agent_tutorials/ai_competitors_analysis_team/fire.py b/ai_agent_tutorials/ai_competitors_analysis_team/fire.py
new file mode 100644
index 0000000..01c4432
--- /dev/null
+++ b/ai_agent_tutorials/ai_competitors_analysis_team/fire.py
@@ -0,0 +1,41 @@
+from phi.agent import Agent
+from phi.tools.firecrawl import FirecrawlTools
+from phi.model.openai import OpenAIChat
+
+firecrawl_tools = FirecrawlTools(
+    api_key="",
+    scrape=False,
+    crawl=True,
+    limit=5
+)
+
+agent = Agent(
+    model=OpenAIChat(id="gpt-4o-mini", api_key=""),
+    tools=[firecrawl_tools],
+    show_tool_calls=True,
+    markdown=True
+)
+
+def extract_competitor_info(competitor_url: str):
+    crawl_response = agent.run(f"Crawl and summarize {competitor_url}")
+    crawled_data = crawl_response.content
+    
+    structured_info = agent.run(
+        f"""Extract the following information from the crawled data:
+        - Product pricing and features - exact pricing numbers from their pricing page
+        - Technology stack information
+        - Marketing messaging/positioning
+        - Customer testimonials/case studies
+
+        Crawled Data:
+        {crawled_data}
+        """
+    )
+    
+    return structured_info.content
+
+if __name__ == "__main__":
+    competitor_url = "https://www.equal.in"
+    competitor_info = extract_competitor_info(competitor_url)
+    print(f"Competitor: {competitor_url}")
+    print(competitor_info)
\ No newline at end of file
diff --git a/ai_agent_tutorials/ai_competitors_analysis_team/firecrawl.py b/ai_agent_tutorials/ai_competitors_analysis_team/firecrawl.py
deleted file mode 100644
index e69de29..0000000
diff --git a/ai_agent_tutorials/ai_competitors_analysis_team/main.py b/ai_agent_tutorials/ai_competitors_analysis_team/main.py
index e69de29..78b9b57 100644
--- a/ai_agent_tutorials/ai_competitors_analysis_team/main.py
+++ b/ai_agent_tutorials/ai_competitors_analysis_team/main.py
@@ -0,0 +1,37 @@
+from firecrawl import FirecrawlApp
+from pydantic import BaseModel, Field
+from typing import Optional
+
+# Define a simple schema for testing
+class SimpleSchema(BaseModel):
+    title: Optional[str] = Field(description="Title of the webpage.")
+    description: Optional[str] = Field(description="Meta description of the webpage.")
+
+# Initialize the FirecrawlApp with your API key
+app = FirecrawlApp(api_key='fc-')  # Replace with your API key
+
+def extract_competitor_info(competitor_url: str):
+
+    try:
+        # Use Firecrawl to scrape and extract data
+        data = app.scrape_url(competitor_url, {
+            'formats': ['extract'],
+            'extract': {
+                'schema': SimpleSchema.model_json_schema(),
+            }
+        })
+        return data.get("extract", {})
+    except Exception as e:
+        return {"error": str(e)}
+
+# Example usage
+if __name__ == "__main__":
+    # Competitor URL to analyze
+    competitor_url = "https://www.equal.in"  # Replace with your competitor URL
+    
+    # Extract competitor information
+    competitor_info = extract_competitor_info(competitor_url)
+    
+    # Print the structured information
+    print(f"Competitor: {competitor_url}")
+    print(competitor_info)
\ No newline at end of file
diff --git a/ai_agent_tutorials/ai_competitors_analysis_team/requirements.txt b/ai_agent_tutorials/ai_competitors_analysis_team/requirements.txt
index 367c38c..e188d2f 100644
--- a/ai_agent_tutorials/ai_competitors_analysis_team/requirements.txt
+++ b/ai_agent_tutorials/ai_competitors_analysis_team/requirements.txt
@@ -1 +1,2 @@
-exa-py==1.0.6   
\ No newline at end of file
+exa-py==1.0.6   
+firecrawl-py==1.9.0
\ No newline at end of file