Upload with huggingface_hub

#agent.py#

@@ -0,0 +1,46 @@
+# + tags=["hide_inp"]
+
+desc = """
+### Gradio Tool
+
+Chain that ask for a command-line question and then runs the bash command. [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/srush/MiniChain/blob/master/examples/bash.ipynb)
+
+(Adapted from LangChain [BashChain](https://langchain.readthedocs.io/en/latest/modules/chains/examples/llm_bash.html))
+"""
+# -
+
+# $
+
+from minichain import show, prompt, OpenAI
+from gradio_tools.tools import StableDiffusionTool, ImageCaptioningTool
+
+class Tool1:
+    description: str = "Tool 1"
+    def run(self, prompt):
+        return prompt
+
+class Tool2:
+    description: str = "Tool 2"
+    def run(self, prompt):
+        return prompt
+
+
+@prompt(OpenAI())
+def gen(model, query):
+    return model(query)
+
+def gradio_example(query):
+    return gen(query)
+
+
+# $
+
+gradio = show(gradio_example,
+              subprompts=[caption],
+              examples=['/home/srush/Projects/MiniChain/examples/63dd90c7-9b8d-4ba4-bc07-a378fd932304/tmph3xi9ylr.jpg', 'Make me a flower'],
+              out_type="markdown",
+              description=desc
+              )
+if __name__ == "__main__":
+    gradio.launch()
+

#base.sh#

@@ -0,0 +1,2 @@
+export OPENAI_API_KEY="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"
+export SERP_KEY="593a073fa4c730efe918e592a538b36e80841bc8f8dd4070c1566920f75ba140"

agent.pmpt.txt

@@ -0,0 +1,36 @@
+Assistant is a large language model trained by OpenAI.
+
+Assistant is designed to be able to assist with a wide range of tasks, from answering simple questions to providing in-depth explanations and discussions on a wide range of topics. As a language model, Assistant is able to generate human-like text based on the input it receives, allowing it to engage in natural-sounding conversations and provide responses that are coherent and relevant to the topic at hand.
+
+Assistant is constantly learning and improving, and its capabilities are constantly evolving. It is able to process and understand large amounts of text, and can use this knowledge to provide accurate and informative responses to a wide range of questions. Additionally, Assistant is able to generate its own text based on the input it receives, allowing it to engage in discussions and provide explanations and descriptions on a wide range of topics.
+
+Overall, Assistant is a powerful tool that can help with a wide range of tasks and provide valuable insights and information on a wide range of topics. Whether you need help with a specific question or just want to have a conversation about a particular topic, Assistant is here to assist.
+
+TOOLS:
+------
+
+Assistant has access to the following tools:
+
+To use a tool, please use the following format:
+
+```
+Thought: Do I need to use a tool? Yes
+Action: the action to take, should be one of [{{tool_names}}]
+Action Input: the input to the action
+Observation: the result of the action
+```
+
+When you have a response to say to the Human, or if you do not need to use a tool, you MUST use the format:
+
+```
+Thought: Do I need to use a tool? No
+AI: [your response here]
+```
+
+Begin!
+
+Previous conversation history:
+{{chat_history}}
+
+New input: {{input}}
+{{agent_scratchpad}}

agent.pmpt.txt~

@@ -0,0 +1,36 @@
+Assistant is a large language model trained by OpenAI.
+
+Assistant is designed to be able to assist with a wide range of tasks, from answering simple questions to providing in-depth explanations and discussions on a wide range of topics. As a language model, Assistant is able to generate human-like text based on the input it receives, allowing it to engage in natural-sounding conversations and provide responses that are coherent and relevant to the topic at hand.
+
+Assistant is constantly learning and improving, and its capabilities are constantly evolving. It is able to process and understand large amounts of text, and can use this knowledge to provide accurate and informative responses to a wide range of questions. Additionally, Assistant is able to generate its own text based on the input it receives, allowing it to engage in discussions and provide explanations and descriptions on a wide range of topics.
+
+Overall, Assistant is a powerful tool that can help with a wide range of tasks and provide valuable insights and information on a wide range of topics. Whether you need help with a specific question or just want to have a conversation about a particular topic, Assistant is here to assist.
+
+TOOLS:
+------
+
+Assistant has access to the following tools:
+
+To use a tool, please use the following format:
+
+```
+Thought: Do I need to use a tool? Yes
+Action: the action to take, should be one of [{{tool_names}}]
+Action Input: the input to the action
+Observation: the result of the action
+```
+
+When you have a response to say to the Human, or if you do not need to use a tool, you MUST use the format:
+
+```
+Thought: Do I need to use a tool? No
+{{ai_prefix}}: [your response here]
+```"""
+
+Begin!
+
+Previous conversation history:
+{chat_history}
+
+New input: {input}
+{agent_scratchpad}

agent.py

@@ -0,0 +1,51 @@
+# + tags=["hide_inp"]
+
+desc = """
+### Gradio Tool
+
+Chain that ask for a command-line question and then runs the bash command. [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/srush/MiniChain/blob/master/examples/bash.ipynb)
+
+(Adapted from LangChain [BashChain](https://langchain.readthedocs.io/en/latest/modules/chains/examples/llm_bash.html))
+"""
+# -
+
+# $
+
+from minichain import show, prompt, OpenAI
+from gradio_tools.tools import StableDiffusionTool, ImageCaptioningTool
+
+class Tool1:
+    description: str = "Tool 1"
+    def run(self, prompt):
+        return prompt
+
+class Tool2:
+    description: str = "Tool 2"
+    def run(self, prompt):
+        return prompt
+
+
+@prompt(StableDiffusionTool())
+def gen(model, query):
+    return model(query)
+
+@prompt(ImageCaptioningTool())
+def caption(model, img_src):
+    return model(img_src)
+
+
+def gradio_example(query):
+    return caption(gen(query))
+
+
+# $
+
+gradio = show(gradio_example,
+              subprompts=[caption],
+              examples=['/home/srush/Projects/MiniChain/examples/63dd90c7-9b8d-4ba4-bc07-a378fd932304/tmph3xi9ylr.jpg', 'Make me a flower'],
+              out_type="markdown",
+              description=desc
+              )
+if __name__ == "__main__":
+    gradio.launch()
+

app.py

@@ -9,15 +9,13 @@ from qa import gradio as qa
 from stats import gradio as stats
 from selfask import gradio as selfask
 from backtrack import gradio as backtrack
+from table import gradio as table
+from gradio_example import gradio as gradio_example
 
 CSS = """
 #clean div.form {border: 0px} 
 #response {border: 0px; background: #ffeec6} 
 #prompt {border: 0px;background: aliceblue} 
-#json {border: 0px} 
-#result {border: 0px; background: #c5e0e5} 
-#inner {margin: 10px; padding: 10px; font-size: 20px; } 
-#inner textarea {border: 0px}
 div.gradio-container {color: black}
 span.head {font-size: 60pt; font-family: cursive;}
 body {
@@ -34,9 +32,9 @@ body {
 with gr.Blocks(css=CSS, theme=gr.themes.Monochrome()) as demo:
     gr.HTML("<center style='background:#B6B7BA'> <span class='head'>Mini</span><img src='https://user-images.githubusercontent.com/35882/227017900-0cacdfb7-37e2-47b1-9347-a233810d3544.png' width='20%' style='display:inline'><span class='head'>Chain</span></center><center> <br><a href='https://github.com/srush/minichain'>[library]</a> </center>")
 
-    gr.TabbedInterface([math_demo, qa,  chat, gatsby,  ner, bash, pal, stats, selfask, backtrack],
-                       ["Math", "QA", "Chat", "Book", "NER", "Bash", "PAL", "Stats", "SelfAsk", "Backtrack"],
+    gr.TabbedInterface([math_demo, qa,  chat, gatsby,  ner, bash, pal, table, gradio_example, stats, selfask, backtrack],
+                       ["Math", "QA", "Chat", "Book", "NER", "Bash", "PAL", "Table", "Gradio", "Stats", "SelfAsk", "Backtrack"],
                        css = CSS)
 
-demo.launch()
+demo.queue().launch()
 

app.py~

@@ -7,15 +7,36 @@ from pal import gradio as pal
 from gatsby import gradio as gatsby
 from qa import gradio as qa
 from stats import gradio as stats
+from selfask import gradio as selfask
+from backtrack import gradio as backtrack
 
-css = "#clean div.form {border: 0px} #response {border: 0px; background: #ffeec6} #prompt {border: 0px;background: aliceblue} #json {border: 0px} #result {border: 0px; background: #c5e0e5} #inner {padding: 20px} #inner textarea {border: 0px} .tabs div.tabitem {border: 0px}"
+CSS = """
+#clean div.form {border: 0px} 
+#response {border: 0px; background: #ffeec6} 
+#prompt {border: 0px;background: aliceblue} 
+#json {border: 0px} 
+#result {border: 0px; background: #c5e0e5} 
+#inner {margin: 10px; padding: 10px; font-size: 20px; } 
+#inner textarea {border: 0px}
+div.gradio-container {color: black}
+span.head {font-size: 60pt; font-family: cursive;}
+body {
+  --text-sm: 12px;
+  --text-md: 16px;
+  --text-lg: 18px;
+  --input-text-size: 16px;
+  --section-text-size: 16px;
+}
+"""
 
-with gr.Blocks(css=css) as demo:
-    gr.HTML("<center> <img width='10%' style='display:inline; padding: 5px' src='https://user-images.githubusercontent.com/35882/218286642-67985b6f-d483-49be-825b-f62b72c469cd.png'> <h1 style='display:inline'> Mini-Chain </h1> <img width='10%' style='display:inline;padding: 5px' src='https://avatars.githubusercontent.com/u/25720743?s=200&v=4'> </center><br><center><a href='https://github.com/srush/minichain'>[code]</a>  <a href='https://user-images.githubusercontent.com/35882/218286642-67985b6f-d483-49be-825b-f62b72c469cd.png'>[docs]</a></center>")
 
-    gr.TabbedInterface([math_demo, qa,  chat, gatsby,  ner, bash, pal, stats],
-                       ["Math", "QA", "Chat", "Book", "NER", "Bash", "PAL", "Stats"],
-                       css = css)
+
+with gr.Blocks(css=CSS, theme=gr.themes.Monochrome()) as demo:
+    gr.HTML("<center style='background:#B6B7BA'> <span class='head'>Mini</span><img src='https://user-images.githubusercontent.com/35882/227017900-0cacdfb7-37e2-47b1-9347-a233810d3544.png' width='20%' style='display:inline'><span class='head'>Chain</span></center><center> <br><a href='https://github.com/srush/minichain'>[library]</a> </center>")
+
+    gr.TabbedInterface([math_demo, qa,  chat, gatsby,  ner, bash, pal, stats, selfask, backtrack],
+                       ["Math", "QA", "Chat", "Book", "NER", "Bash", "PAL", "Stats", "SelfAsk", "Backtrack"],
+                       css = CSS)
 
 demo.launch()
 

backtrack.py

@@ -11,11 +11,15 @@ Chain that backtracks on failure. [![Open In Colab](https://colab.research.googl
 from minichain import prompt, Mock, show, OpenAI
 import minichain
 
-@prompt(Mock(["dog", "blue", "cat"]))
+@prompt(Mock(["dog", "blue", "cat"]), stream=True)
 def prompt_generation(model):
-    return model("")
+    out = ""
+    for token in model.stream(""):
+        out += token
+        yield out
+    yield out
 
-@prompt(OpenAI(), template="Answer 'yes' is {{query}} is a color. Answer:")
+@prompt(Mock(["No", "Yes"]), template="Answer 'yes' is {{query}} is a color. Answer:", stream=False)
 def prompt_validation(model, x):
     out = model(dict(query=x))
     if out.strip().lower().startswith("yes"):
@@ -35,4 +39,4 @@ gradio = show(run,
             )
 
 if __name__ == "__main__":
-    gradio.launch()
+    gradio.queue().launch()

backtrack.py~

@@ -2,33 +2,37 @@
 desc = """
 ### Backtrack on Failure
 
-Chain that backtracks on failure. [[Code](https://github.com/srush/MiniChain/blob/main/examples/backtrack.py)]
+Chain that backtracks on failure. [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/srush/MiniChain/blob/master/examples/backtrack.ipynb)
 
 """
 # -
 
-
-from minichain import prompt, Mock, show
+# $
+from minichain import prompt, Mock, show, OpenAI
 import minichain
 
-@prompt(Mock(["red", "blue"]))
-def prompt_function1(model, x):
-    return model(x)
+@prompt(Mock(["dog", "blue", "cat"]))
+def prompt_generation(model):
+    return model("")
 
-@prompt(Mock(["b"]), template_file="test.pmpt.tpl")
-def prompt_function2(model, x):
-    if x == "red":
-        return model.fail(1)
-    return model(dict(x=x))
+@prompt(OpenAI(), template="Answer 'yes' is {{query}} is a color. Answer:")
+def prompt_validation(model, x):
+    out = model(dict(query=x))
+    if out.strip().lower().startswith("yes"):
+        return x
+    return model.fail(1)
     
-def run(query):
-    x = prompt_function1(query)
-    return prompt_function2(prompt_function2(x))
-
-
-demo = show(run,
-            examples=["a"],
-            subprompts=[prompt_function1, prompt_function2, prompt_function2])
+def run():
+    x = prompt_generation()
+    return prompt_validation(x)
+# $
+
+gradio = show(run,
+            examples = [],
+            subprompts=[prompt_generation, prompt_validation],
+            code=open("backtrack.py", "r").read().split("$")[1].strip().strip("#").strip(),
+            out_type="markdown"
+            )
 
 if __name__ == "__main__":
-    demo.launch()
+    gradio.launch()

base.sh

@@ -0,0 +1,2 @@
+export OPENAI_API_KEY="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"
+export SERP_KEY="593a073fa4c730efe918e592a538b36e80841bc8f8dd4070c1566920f75ba140"

bash.py

@@ -39,5 +39,5 @@ gradio = show(bash,
               code=open("bash.py", "r").read().split("$")[1].strip().strip("#").strip(),
               )
 if __name__ == "__main__":
-    gradio.launch()
+    gradio.queue().launch()
 

chat.py

@@ -29,6 +29,9 @@ class State:
         memory = self.memory if len(self.memory) < MEMORY else self.memory[1:]
         return State(memory + [(self.human_input, response)])
 
+    def __str__(self):
+        return self.memory[-1][-1]
+
 # Chat prompt with memory
 
 @prompt(OpenAI(), template_file="chat.pmpt.tpl")
@@ -58,6 +61,6 @@ gradio = show(lambda command, state: chat_prompt(replace(state, human_input=comm
               code=open("chat.py", "r").read().split("$")[1].strip().strip("#").strip(),
 )
 if __name__ == "__main__":
-    gradio.launch()
+    gradio.queue().launch()
 
 

gatsby.py

@@ -19,7 +19,7 @@ from minichain import prompt, show, HuggingFaceEmbed, OpenAI
 gatsby = datasets.load_from_disk("gatsby")
 gatsby.add_faiss_index("embeddings")
 
-# Fast KNN retieval prompt
+# Fast KNN retrieval prompt
 
 @prompt(HuggingFaceEmbed("sentence-transformers/all-mpnet-base-v2"))
 def get_neighbors(model, inp, k=1):
@@ -32,7 +32,7 @@ def get_neighbors(model, inp, k=1):
 def ask(model, query, neighbors):
     return model(dict(question=query, docs=neighbors))
 
-def gatsby(query):
+def gatsby_q(query):
     n = get_neighbors(query)
     return ask(query, n)
 
@@ -40,7 +40,7 @@ def gatsby(query):
 # $
 
 
-gradio = show(gatsby,
+gradio = show(gatsby_q,
               subprompts=[get_neighbors, ask],
               examples=["What did Gatsby do before he met Daisy?",
                         "What did the narrator do after getting back to Chicago?"],
@@ -49,4 +49,4 @@ gradio = show(gatsby,
               code=open("gatsby.py", "r").read().split("$")[1].strip().strip("#").strip()
               )
 if __name__ == "__main__":
-    gradio.launch()
+    gradio.queue().launch()

gradio_example.py

@@ -0,0 +1,51 @@
+# + tags=["hide_inp"]
+
+desc = """
+### Gradio Tool
+
+Examples using the gradio tool [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/srush/MiniChain/blob/master/examples/gradio_example.ipynb)
+
+"""
+# -
+
+# $
+
+from minichain import show, prompt, OpenAI, OpenAIStream
+import gradio as gr
+from gradio_tools.tools import StableDiffusionTool, ImageCaptioningTool
+
+@prompt(OpenAIStream(), stream=True)
+def picture(model, query):
+    out = ""
+    for r in model.stream(query):
+        out += r
+        yield out
+
+@prompt(StableDiffusionTool(), stream=True, block_input=lambda: gr.Textbox(label=""))
+def gen(model, query):
+    for r in model.stream(query):
+        yield "https://htmlcolorcodes.com/assets/images/colors/baby-blue-color-solid-background-1920x1080.png"
+    yield r
+
+@prompt(ImageCaptioningTool(), block_output=lambda: gr.Textbox(label=""))
+def caption(model, img_src):
+    return model(img_src)
+
+def gradio_example(query):
+    return caption(gen(picture(query)))
+
+
+# $
+
+gradio = show(gradio_example,
+              subprompts=[picture, gen, caption],
+              examples=['Describe a one-sentence fantasy scene.',
+                        'Describe a one-sentence scene happening on the moon.'],
+              out_type="markdown",
+              description=desc,
+              css="#advanced {display: none}"
+
+              )
+if __name__ == "__main__":
+    gradio.queue().launch()
+

gradio_example.py~

@@ -0,0 +1,41 @@
+# + tags=["hide_inp"]
+
+desc = """
+### Gradio Tool
+
+Chain that ask for a command-line question and then runs the bash command. [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/srush/MiniChain/blob/master/examples/bash.ipynb)
+
+(Adapted from LangChain [BashChain](https://langchain.readthedocs.io/en/latest/modules/chains/examples/llm_bash.html))
+"""
+# -
+
+# $
+
+from minichain import show, prompt, OpenAI
+from gradio_tools.tools import StableDiffusionTool, ImageCaptioningTool
+
+
+@prompt(StableDiffusionTool())
+def gen(model, query):
+    return model(query)
+
+@prompt(ImageCaptioningTool())
+def caption(model, img_src):
+    return model(img_src)
+
+
+def gradio_example(query):
+    return caption(gen(query))
+
+
+# $
+
+gradio = show(gradio_example,
+              subprompts=[caption],
+              examples=['/home/srush/Projects/MiniChain/examples/63dd90c7-9b8d-4ba4-bc07-a378fd932304/tmph3xi9ylr.jpg', 'Make me a flower'],
+              out_type="markdown",
+              description=desc
+              )
+if __name__ == "__main__":
+    gradio.launch()
+

gradio_tool.py~

@@ -0,0 +1,43 @@
+# + tags=["hide_inp"]
+
+desc = """
+### Gradio Tool
+
+Chain that ask for a command-line question and then runs the bash command. [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/srush/MiniChain/blob/master/examples/bash.ipynb)
+
+(Adapted from LangChain [BashChain](https://langchain.readthedocs.io/en/latest/modules/chains/examples/llm_bash.html))
+"""
+# -
+
+# $
+
+from minichain import show, prompt, OpenAI
+
+
+@prompt(OpenAI(), template_file = "bash.pmpt.tpl")
+def output_prompt(model, query):
+    x = model(dict(question=query))
+    return "\n".join(x.strip().split("\n")[1:-1])
+
+@prompt(Bash())
+def bash_run(model, x):
+    return model(x)
+
+def gradio_tool(query):
+    return bash_run(cli_prompt(query))
+
+
+# $
+
+gradio = show(bash,
+              subprompts=[cli_prompt, bash_run],
+              examples=['Go up one directory, and then into the minichain directory,'
+                        'and list the files in the directory',
+                        "Please write a bash script that prints 'Hello World' to the console."],
+              out_type="markdown",
+              description=desc,
+              code=open("bash.py", "r").read().split("$")[1].strip().strip("#").strip(),
+              )
+if __name__ == "__main__":
+    gradio.launch()
+

math_demo.py

@@ -41,6 +41,6 @@ gradio = show(math_demo,
               code=open("math_demo.py", "r").read().split("$")[1].strip().strip("#").strip(),
               )
 if __name__ == "__main__":
-    gradio.launch()
+    gradio.queue().launch()
 # -
 

ner.py

@@ -39,4 +39,4 @@ gradio = show(ner,
               )
 
 if __name__ == "__main__":
-    gradio.launch()
+    gradio.queue().launch()

pal.py

@@ -8,14 +8,19 @@ Chain for answering complex problems by code generation and execution. [![Open I
 
 # $
 
-from minichain import prompt, show, OpenAI, Python
+from minichain import prompt, show, OpenAIStream, Python
+import gradio as gr
 
-@prompt(OpenAI(), template_file="pal.pmpt.tpl")
+@prompt(OpenAIStream(), template_file="pal.pmpt.tpl", stream=True)
 def pal_prompt(model, question):
-    return model(dict(question=question))
+    out = ""
+    for t in model.stream(dict(question=question)):
+        out += t
+        yield out
 
-@prompt(Python())
+@prompt(Python(), block_input=lambda: gr.Code(language="python"))
 def python(model, inp):
+    print(inp)
     return float(model(inp + "\nprint(solution())"))
 
 def pal(question):
@@ -37,4 +42,4 @@ gradio = show(pal,
               )
 
 if __name__ == "__main__":
-    gradio.launch()
+    gradio.queue().launch()

qa.py

@@ -55,5 +55,5 @@ gradio = show(qa,
               code=open("qa.py", "r").read().split("$")[1].strip().strip("#").strip(),
               )
 if __name__ == "__main__":
-    gradio.launch()
+    gradio.queue().launch()
 

selfask.py

@@ -59,6 +59,6 @@ gradio = show(selfask,
               out_type="json"
               )
 if __name__ == "__main__":
-    gradio.launch()
+    gradio.queue().launch()
 
 

stats.py

@@ -50,7 +50,7 @@ gradio = show(lambda passage: stats(passage),
               code=open("stats.py", "r").read().split("$")[1].strip().strip("#").strip(),
 )
 if __name__ == "__main__":
-    gradio.launch()
+    gradio.queue().launch()
 
 
 # ExtractionPrompt().show({"passage": "Harden had 10 rebounds."},

table.pmpt.txt

@@ -0,0 +1,31 @@
+You are a utility built to extract structured information from documents. You are returning a TSV table. Here are the headers .
+
+----
+{{type}}        {% for k in player_keys %}{{k[0]}}{{"\t" if not loop.last}}{% endfor %}
+----
+
+Return the rest of the table in TSV format. Here are some examples
+
+{% for example in examples %}
+Example
+---
+{{example.input}}
+---
+
+Output
+---
+{{example.output}}
+---
+{% endfor %}
+
+Article:
+----
+{{passage}}
+----
+
+All other values should be numbers or _.
+Only include numbers that appear explicitly in the passage below.
+If you cannot find the value in the table, output _. Most cells will be _.
+
+Ok, here is the correctly valid TSV with headers and nothing else. Remember only include values that are directly written in the article. Do not guess or combine rows.
+

table.py

@@ -0,0 +1,83 @@
+# + tags=["hide_inp"]
+desc = """
+### Table
+
+Example of extracting tables from a textual document. [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/srush/MiniChain/blob/master/examples/table.ipynb)
+
+"""
+# -
+
+# $
+import pandas as pd
+from minichain import prompt, Mock, show, OpenAIStream
+import minichain
+import json
+import gradio as gr
+
+rotowire = json.load(open("data.json"))
+names = {
+    '3-pointer percentage': 'FG3_PCT',
+    '3-pointers attempted': 'FG3A',
+    '3-pointers made': 'FG3M',
+    'Assists': 'AST',
+    'Blocks': 'BLK',
+    'Field goal percentage': 'FG_PCT',
+    'Field goals attempted': 'FGA',
+    'Field goals made': 'FGM',
+    'Free throw percentage': 'FT_PCT',
+    'Free throws attempted': 'FTA',
+    'Free throws made': 'FTM',
+    'Minutes played': 'MIN',
+    'Personal fouls': 'PF',
+    'Points': 'PTS',
+    'Rebounds': 'REB',
+    'Rebounds (Defensive)': 'DREB',
+    'Rebounds (Offensive)': 'OREB',
+    'Steals': 'STL',
+    'Turnovers': 'TO'
+}
+# Convert an example to dataframe
+def to_df(d):
+    players = {player for v in d.values() if v is not None for player, _  in v.items()}
+    lookup = {k: {a: b for a, b in v.items()} for k,v in d.items()}
+    rows = [{"player": p} | {k: "_" if p not in lookup.get(k, []) else lookup[k][p] for k in names.keys()}
+            for p in players]
+    return pd.DataFrame.from_dict(rows).astype("str").sort_values(axis=0, by="player", ignore_index=True).transpose()
+
+
+# Make few shot examples
+few_shot_examples = 2
+examples = []
+for i in range(few_shot_examples):
+    examples.append({"input": rotowire[i][1],
+                     "output": to_df(rotowire[i][0][1]).transpose().set_index("player").to_csv(sep="\t")})
+
+@prompt(OpenAIStream(),
+        template_file="table.pmpt.txt",
+        block_output=gr.HTML,
+        stream=True)
+def extract(model, passage, typ):
+    state = []
+    out = ""
+    for token in model.stream(dict(player_keys=names.items(), examples=examples, passage=passage, type=typ)):
+        out += token
+        html = "<table><tr><td>" + out.replace("\t", "</td><td>").replace("\n", "</td></tr><tr><td>")  + "</td></td></table>"
+        yield html
+    yield html
+
+
+
+def run(query):
+    return extract(query, "Player")
+
+# $
+
+gradio = show(run,
+              examples = [rotowire[i][1] for i in range(50, 55)],
+              subprompts=[extract],
+              code=open("table.py", "r").read().split("$")[1].strip().strip("#").strip(),
+              out_type="markdown"
+            )
+
+if __name__ == "__main__":
+    gradio.queue().launch()

table.py~

@@ -0,0 +1,76 @@
+# + tags=["hide_inp"]
+desc = """
+### Table
+
+"""
+# -
+
+# $
+import pandas as pd
+from minichain import prompt, Mock, show, OpenAIStream
+import minichain
+
+rotowire = json.load(open("data.json"))
+names = {
+    '3-pointer percentage': 'FG3_PCT',
+    '3-pointers attempted': 'FG3A', 
+    '3-pointers made': 'FG3M',
+    'Assists': 'AST',
+    'Blocks': 'BLK',
+    'Field goal percentage': 'FG_PCT',
+    'Field goals attempted': 'FGA',  
+    'Field goals made': 'FGM',
+    'Free throw percentage': 'FT_PCT',
+    'Free throws attempted': 'FTA', 
+    'Free throws made': 'FTM', 
+    'Minutes played': 'MIN',
+    'Personal fouls': 'PF',
+    'Points': 'PTS',
+    'Rebounds': 'REB',
+    'Rebounds (Defensive)': 'DREB',
+    'Rebounds (Offensive)': 'OREB', 
+    'Steals': 'STL',
+    'Turnovers': 'TO'
+}
+# Convert an example to dataframe
+def to_df(d):
+    players = {player for v in d.values() if v is not None for player, _  in v.items()}
+    lookup = {k: {a: b for a, b in v.items()} for k,v in d.items()}
+    rows = [{"player": p} | {k: "_" if p not in lookup.get(k, []) else lookup[k][p] for k in names.keys()}
+            for p in players]
+    return pd.DataFrame.from_dict(rows).astype("str").sort_values(axis=0, by="player", ignore_index=True).transpose()
+
+
+# Make few shot examples
+few_shot_examples = 2
+examples = []
+for i in range(few_shot_examples):
+    examples.append({"input": rotowire[i][1], 
+                     "output": to_df(rotowire[i][0][1]).transpose().set_index("player").to_csv(sep="\t")})
+
+@prompt(OpenAIStream(), template_file="table.pmpt.txt", stream=True)
+def extract(model, passage, typ):
+    state = []
+    out = ""
+    for token in model(dict(player_keys=names.items(), examples=examples, passage=passage, type=typ)):
+        out += token
+        html = "<table><tr><td>" + out.replace("\t", "</td><td>").replace("\n", "</td></tr><tr><td>")  + "</td></td></table>"
+        yield html
+    yield html
+        
+
+    
+def run(query):
+    return extract(query)
+
+# $
+
+gradio = show(run,
+              gr.Examples([rotowire[i][1] for i in range(50, 55)], inputs=[passage])
+              subprompts=[prompt_generation, prompt_validation],
+              code=open("table.py", "r").read().split("$")[1].strip().strip("#").strip(),
+              out_type="markdown"
+            )
+
+if __name__ == "__main__":
+    gradio.queue().launch()