Modfiy README.md / add gif

2 files changed, 47 insertions, 76 deletions

diff --git a/README.md b/README.md
index 653c2dc..175a386 100644
--- a/README.md
+++ b/README.md
@@ -1,84 +1,47 @@
-gpt-chat-cli
-------------
+gpt-chat-cli: a simple yet powerful ChatGPT CLI
+-----------------------------------------------
 
-> This is a WIP. Please consider it experimental.
+## Intro
 
-A simple yet powerful ChatGPT CLI.
+`gpt-chat-cli` is a simple, general purpose ChatGPT CLI. It brings the power of ChatGPT to the command line. It aims to be easy to use and configurable.
+
+Some of the features include:
+- Streaming, real-time output
+- Interactive sessions with color and adornments
+- Support for any model which can be called through OpenAI's chat completions API.
+  [See model endpoint compatibility.](https://platform.openai.com/docs/models/model-endpoint-compatibility)
+- Modify model parameters including temperature, frequency penalty, presence penalty, top p, and the maximum number of tokens emitted
+- Dynamic code syntax highlighting
+- List available models
+- Respects unix norms. Input can be gathered from pipes, heredoc, files, and arbitrary file descriptors.
+
+![gpt-cli](./assets/images/gpt.gif)
 
 ### Installation:
 
-```
+```bash
 pip install gpt-chat-cli
 ```
 
-The OpenAI API uses API keys for authentication. Visit your [API Keys page](https://platform.openai.com/account/api-keys) to retrieve the API key you'll use in your requests. Then, source the `OPENAI_API_KEY` environmental variable in your shell's configuration file. (That is, `~/.bashrc` or `~/.zshrc` for the Bash or Zsh shell, respectively.)
-```
+The OpenAI API uses API keys for authentication. Visit your [API Keys page](https://platform.openai.com/account/api-keys) to retrieve the API key you'll use in your requests. 
+```bash
 export OPENAI_API_KEY="INSERT_SECRET_KEY"
 ```
 
-### Examples:
-
-```
-[#] gpt-chat-cli "Tell me about Joseph Weizenbaum"
-
-[gpt-3.5-turbo-0301] Joseph Weizenbaum (1923-2008) was a German-American computer scientist and philosopher who is best known for creating the ELIZA program, one of the earliest examples of natural language processing. ELIZA was a computer program that simulated a psychotherapist and could engage in a conversation with a human user by using simple pattern-matching techniques to respond to the user's input.
-
-Weizenbaum was born in Berlin, Germany, and fled the country with his family in 1936 to escape Nazi persecution. He later studied mathematics and physics at Wayne State University in Detroit, Michigan, and received a Ph.D. in computer science from the Massachusetts Institute of Technology (MIT) in 1956.
-
-In addition to his work on ELIZA, Weizenbaum was also a prominent critic of artificial intelligence and the use of computers in society. He argued that computers could not truly understand human language or emotions, and that the increasing reliance on technology could lead to dehumanization and loss of personal autonomy.
+Then, source the `OPENAI_API_KEY` environmental variable in your shell's configuration file. (That is, `~/.bashrc` or `~/.zshrc` for the Bash or Zsh shell, respectively.)
 
-Weizenbaum's writings and lectures on the social and ethical implications of technology have had a lasting impact on the field of computer science and continue to be studied and debated today.
+```bash
+source ~/.bashrc
 ```
 
-```
-[#] gpt-chat-cli -n 2 "Write Rust code to find the variance of a list of floating point numbers"
-[gpt-3.5-turbo-0301 1/2] Here's the Rust code to find the variance of a list of floating point numbers:
-
-fn variance(numbers: &[f64]) -> Option<f64> {
-    let n = numbers.len() as f64;
-    if n <= 1.0 {
-        return None;
-    }
-    let mean = numbers.iter().sum::<f64>() / n;
-    let variance = numbers.iter().map(|x| (x - mean).powi(2)).sum::<f64>() / (n - 1.0);
-    Some(variance)
-}
-
-This function takes a slice of floating point numbers as input and returns an `Option<f64>` which is the variance of the numbers. If the length of the slice is less than or equal to 1, the function returns `None` since variance cannot be calculated for such cases.
+### User guide
 
-The function calculates the mean of the numbers by summing up all the numbers and dividing by the length of the slice. It then calculates the variance by iterating over the numbers, subtracting the mean from each number, squaring the result and summing up all the squares. Finally, it divides the sum of squares by (n-1) to get the variance.
-
-Note that we use `powi(2)` instead of `powf(2.0)` to avoid floating point errors.
-
-[gpt-3.5-turbo-0301 2/2] Here's one possible implementation:
-
-fn variance(data: &[f64]) -> Option<f64> {
-    let n = data.len() as f64;
-    if n <= 1.0 {
-        return None; // need at least 2 data points
-    }
-    let mean = data.iter().sum::<f64>() / n;
-    let variance = data.iter().map(|x| (x - mean).powi(2)).sum::<f64>() / (n - 1.0);
-    Some(variance)
-}
-
-This function takes a slice of `f64` values as input, and returns an `Option<f64>` which is `None` if the input slice has fewer than 2 elements (since we need at least 2 data points to compute a variance), or the computed variance otherwise.
-
-The variance is computed using the formula:
-
-variance = sum((x - mean)^2) / (n - 1)
-
-where `x` is each data point, `mean` is the mean of the data, `n` is the number of data points, and `^2` means "squared". The `map()` method is used to compute the squared differences `(x - mean)^2`, and the `sum()` method is used to add them up. Finally, the result is divided by `n - 1` to get the unbiased sample variance (which is what we usually want).
-
-Note that this implementation uses the `powi()` method to compute the squared differences, which is faster than using the `powf()` method with a constant exponent of 2.
-```
-
-### Usage:
+#### Usage:
 
 ```
-usage: gcli.py [-h] [-m MODEL] [-t TEMPERATURE] [-f FREQUENCY_PENALTY] [-p PRESENCE_PENALTY] [-k MAX_TOKENS] [-s TOP_P] [-n N_COMPLETIONS] [--adornments {on,off,auto}] [--color {on,off,auto}] [--version] [-l]
-               [-i]
-               [message]
+usage: gpt-chat-cli [-h] [-m MODEL] [-t TEMPERATURE] [-f FREQUENCY_PENALTY] [-p PRESENCE_PENALTY] [-k MAX_TOKENS] [-s TOP_P] [-n N_COMPLETIONS] [--system-message SYSTEM_MESSAGE] [--adornments {on,off,auto}]
+                    [--color {on,off,auto}] [--version] [-l] [-i] [--prompt-from-fd PROMPT_FROM_FD | --prompt-from-file PROMPT_FROM_FILE]
+                    [message]
 
 positional arguments:
   message               The contents of the message. When in a interactive session, this is the initial prompt provided.
@@ -100,6 +63,8 @@ options:
                         comprising the top 10% probability mass are considered.
   -n N_COMPLETIONS, --n-completions N_COMPLETIONS
                         How many chat completion choices to generate for each input message.
+  --system-message SYSTEM_MESSAGE
+                        Specify an alternative system message.
   --adornments {on,off,auto}
                         Show adornments to indicate the model and response. Can be set to 'on', 'off', or 'auto'.
   --color {on,off,auto}
@@ -107,17 +72,23 @@ options:
   --version             Print version and exit
   -l, --list-models     List models and exit
   -i, --interactive     Start an interactive session
+  --prompt-from-fd PROMPT_FROM_FD
+                        Obtain the initial prompt from the specified file descriptor
+  --prompt-from-file PROMPT_FROM_FILE
+                        Obtain the initial prompt from the specified file
 ```
 
-### Features:
-
-- [x] Streaming real-time output
-- [x] Color and adornments
-- [x] Support for multiple completions
-- [x] Support for any model which is supported through the chat completions API. [See model endpoint compatibility.](https://platform.openai.com/docs/models/model-endpoint-compatibility)
-- [x] Capability to modify parameters including temperature, frequency penalty, presence penalty, top p, and the maximum number of tokens emitted
-- [x] Interactive sessions
-- [x] Code syntax highlighting (when the LLM labels code blocks)
-- [x] Interactive mode with emacs-style editing
-- [x] Input from standard in, pipes, and heredoc
-- [x] List available models
+#### Environmental Variables
+
+Environmental variables can control default model parameters. They are overwritten by command-line parameters if specified.
+
+| Environmental Variable | Controls | Default Value |
+| --- | --- | --- |
+| `GPT_CLI_MODEL` | ID of the model to use | "gpt-3.5-turbo" |
+| `GPT_CLI_TEMPERATURE` | Sampling temperature to use, between 0 and 2 | 0.5 |
+| `GPT_CLI_FREQUENCY_PENALTY` | Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far | 0 |
+| `GPT_CLI_PRESENCE_PENALTY` | Number between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far | 0 |
+| `GPT_CLI_MAX_TOKENS` | The maximum number of tokens to generate in the chat completion | 2048 |
+| `GPT_CLI_TOP_P` | An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass | 1 |
+| `GPT_CLI_N_COMPLETIONS` | How many chat completion choices to generate for each input message | 1 |
+| `GPT_CLI_SYSTEM_MESSAGE` | Specify an alternative system message | See |
diff --git a/assets/images/gpt.gif b/assets/images/gpt.gif
new file mode 100644
index 0000000..70d70c4
--- /dev/null
+++ b/assets/images/gpt.gif